{"slug": "gemini-cli", "name": "Gemini CLI", "category": "ai-coding", "category_name": "AI coding tools", "description": "Google's open-source terminal AI coding agent powered by Gemini models.", "recommenders": 13, "commenters": 54, "warners": 27, "score": -8.86, "rising": -1.6, "first": "2025-11-13", "latest": "2026-09-21", "why": "A minority of HN commenters like it as a capable coding agent and for explaining code, and some praise its plain terminal interface and open-source extensibility.The most common complaint is that it trails Claude Code and Codex, with frequent bugs, infinite loops and weak tool calling, and some note it is being phased out for Antigravity CLI.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 13, "neutral": 17, "compare": 0, "warn": 26}, "spark": [4, 2, 3, 0, 1, 0, 0, 1, 0, 1, 1, 0], "listed": true, "loved": ["Coding capability", "Terminal interface"], "caveat": "Behind competitors", "short_quote": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage.", "icon": true, "domain": null, "github_repo": "github.com/google-gemini/gemini-cli", "official_url": "https://github.com/google-gemini/gemini-cli", "pricing": "open source", "aliases": ["Gemini CLI", "gemini-cli"], "verdict": [{"text": "A minority of HN commenters like it as a capable coding agent and for explaining code,", "cites": [{"id": 788680097, "item_id": 49292506, "url": "https://news.ycombinator.com/item?id=49292506", "by": "gazebo2", "date": "2026-08-13", "stance": "recommend", "quote": "found it also an effective coding agent", "excerpt": "I similarly have a weird affinity for Gemini that I can't really articulate. I used Gemini's free chat and found it great for exploring technical topics (and random one-off general walking-around-questions) and appreciated its speed, tone and accuracy. I spent a month playing with Gemini CLI / Antigravity and found it also an effective coding agent, at least for my workflow (entirely in the loop development and revie", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Gemini 3.7 Flash"}, {"id": 748116898, "item_id": 46757306, "url": "https://news.ycombinator.com/item?id=46757306", "by": "lywald", "date": "2026-01-25", "stance": "recommend", "quote": "Claude was down so I used Gemini CLI and it turns out it is quite capable also.", "excerpt": "Hey HN, Pierre here. I\u2019m the solo creator of EbookForge. I know the skepticism around AI-generated content (and \"slop\") is high right now. I share it. But I built this tool because I found that for self-education and structuring ideas, LLMs are actually incredible. I used EbookForge itself to generate a massive, 150-page catalog containing 1,000+ book concepts. This goes beyond concept: we can then recursively genera", "context": "solo creator", "use_case": "vibe-coding", "switched_from": "Claude", "story": "Show HN: I used my book generator to generate a catalog of books it can generate"}, {"id": 736169824, "item_id": 46010614, "url": "https://news.ycombinator.com/item?id=46010614", "by": "keithba", "date": "2025-11-21", "stance": "recommend", "quote": "I used Gemini CLI (once it came, AI Studio before that) with Gemini 2.5 Pro to create one with Rust and LLVM", "excerpt": "LLMs can definitely write programming languages. I used Gemini CLI (once it came, AI Studio before that) with Gemini 2.5 Pro to create one with Rust and LLVM: https://github.com/GoogleCloudPlatform/aether This was completely vibe coded - I never had to edit the code, though it was very interactive. The whole thing tool less than a month of some evenings and weekends. (Note: it\u2019s ugly on purpose, as I\u2019m playing with i", "context": "", "use_case": "vibe coding a language", "switched_from": "AI Studio", "story": "FAWK: LLMs can write a language interpreter"}, {"id": 795435617, "item_id": 49714726, "url": "https://news.ycombinator.com/item?id=49714726", "by": "pid0x17", "date": "2026-09-15", "stance": "recommend", "quote": "I use to love using Gemini CLI to have Gemini explain a module for me", "excerpt": "We moved from Gemini to Claude recently, and while I absolutely love Claude's coding capabilities, I often find myself extremely overwhelmed whenever I ask Claude to explain something to me, whereas that was a breeze with Gemini Pro models. It generates pages for what could be explained in 1-2 paragraphs. I use to love using Gemini CLI to have Gemini explain a module for me before I dive into the code and point out s", "context": "team moved from Gemini to Claude", "use_case": "explaining code modules", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}]}, {"text": " and some praise its plain terminal interface and open-source extensibility.", "cites": [{"id": 757442993, "item_id": 47340187, "url": "https://news.ycombinator.com/item?id=47340187", "by": "xrd", "date": "2026-03-11", "stance": "recommend", "quote": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage.", "excerpt": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage. Why are they great? Because it is simply text that I use to interact with them. That's really simple and powerful. I don't understand why I can't levitate that simple interface into a web UI inside my phone browser? It feels like this should be as simple as webmux (tmux on the web). But it feels surprisingly elusive. I would really like som", "context": "", "use_case": "terminal coding agent", "switched_from": "", "story": "Karpathy is searching for the Agentic IDE"}, {"id": 735788498, "item_id": 45986781, "url": "https://news.ycombinator.com/item?id=45986781", "by": "freediver", "date": "2025-11-19", "stance": "warn", "quote": "Gemini CLI was never a contender (between log in/activation/rate limits - wth), will say though that Gemini CLI has the nicest terminal UI.", "excerpt": "First time that there is a worthy alternative to Claude Code. Codex Max solved a problem I had Claude Code fail multiple times. Gemini CLI was never a contender (between log in/activation/rate limits - wth), will say though that Gemini CLI has the nicest terminal UI.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Building more with GPT-5.1-Codex-Max"}, {"id": 737029072, "item_id": 46064317, "url": "https://news.ycombinator.com/item?id=46064317", "by": "terrywang", "date": "2025-11-26", "stance": "recommend", "quote": "Gemini CLI is open source and being actively developed, which is cool", "excerpt": "Gemini CLI at this stage isn't good at complex coding tasks (vs. Claude Code, Codex, Cursor CLI, Qoder CLI, etc.). Mostly because of the simple ReAct loop, compounded by relatively weak tool calling capability of the Gemini 2.5 Pro model. > I haven't tried complex coding tasks using Gemini 3.0 Pro Preview yet. I reckon it won't be materially different. Gemini CLI is open source and being actively developed, which is ", "context": "", "use_case": "orchestrating tasks and workflows", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}]}, {"text": "The most common complaint is that it trails Claude Code and Codex,", "cites": [{"id": 753690304, "item_id": 47105644, "url": "https://news.ycombinator.com/item?id=47105644", "by": "dudeinhawaii", "date": "2026-02-21", "stance": "warn", "quote": "I find that Gemini CLI is still considerably worse than both Codex and Claude Code.", "excerpt": "After 2 days of giving it a go, I find that Gemini CLI is still considerably worse than both Codex and Claude Code. The model itself also has strange behaviors that seem like it gets randomly replaced with Gemini-3-Flash or something else. I'll explain. Once agentic coding was a bust, I gave it a run as a daily driver for AI assistant. It performed fairly well but then began behaving strangely. It would lose context ", "context": "Google AI Ultra subscriber", "use_case": "agentic coding", "switched_from": "", "story": "Gemini 3.1 Pro Preview"}, {"id": 759506163, "item_id": 47469135, "url": "https://news.ycombinator.com/item?id=47469135", "by": "arthurjean", "date": "2026-03-21", "stance": "warn", "quote": "well above Codex, Gemini CLI or Mistral Vibe in my experience", "excerpt": "I've used both. I stuck with Claude Code, the ergonomics are better and the internals are clearly optimized for Opus which I use daily, you can feel it. That said OpenCode is still a very good alternative, well above Codex, Gemini CLI or Mistral Vibe in my experience.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "OpenCode \u2013 The open source AI coding agent"}, {"id": 744514739, "item_id": 46532171, "url": "https://news.ycombinator.com/item?id=46532171", "by": "vedmakk", "date": "2026-01-07", "stance": "warn", "quote": "Gemini CLI is a pain to use - after getting used to the smoothness of Claude\nCode.", "excerpt": "I don't get the Gemini 3 hype... yes it's their first usable model, but its not even close to what Opus 4.5 and GPT 5.2 can do. Maybe on Benchmarks... but I'm forced to use Gemini at work everyday, while I use Opus 4.5 / GPT 5.2 privately every day... and Gemini is just lacking so much wit, creativity and multi-step problem solving skills compared to Opus. Not to mention that Gemini CLI is a pain to use - after getti", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Google Got Its Groove Back and Edged Ahead of OpenAI"}, {"id": 737045889, "item_id": 46065368, "url": "https://news.ycombinator.com/item?id=46065368", "by": "MarkMarine", "date": "2025-11-26", "stance": "warn", "quote": "I tried Gemini cli for about 5 seconds and was so frustrated, it\u2019s so stupid at navigation in the codebase", "excerpt": "I love the model, hate the tool. I\u2019ve taken complex stuff and given it to Gemini 3 and been impressed, but Anthropic has the killer app with Claude Code. The interplay of sonnet (a decent model) and the tools and workflow they\u2019ve got with Claude code around it supercharge the outcome. I tried Gemini cli for about 5 seconds and was so frustrated, it\u2019s so stupid at navigation in the codebase it takes 10x as long to do ", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}]}, {"text": " with frequent bugs, infinite loops and weak tool calling,", "cites": [{"id": 786995440, "item_id": 49187215, "url": "https://news.ycombinator.com/item?id=49187215", "by": "pixel_popping", "date": "2026-08-05", "stance": "warn", "quote": "Gemini CLI was utter shit and always had delayed model release on top of it.", "excerpt": "Gemini CLI was utter shit and always had delayed model release on top of it. I'm not kidding when 1 hour automated vibe-coding led to an agent 30% cheaper and faster on 1:1 tasks with the SAME Gemini models (re-ran on 50 past tasks), genuinely shameful that Google even released this, it was so buggy, trapped in infinite loops as well and so-on. Antigravity is a whole different product, the only way I use it currently", "context": "", "use_case": "agentic coding CLI", "switched_from": "", "story": "Antigravity 2.0"}, {"id": 747812609, "item_id": 46738288, "url": "https://news.ycombinator.com/item?id=46738288", "by": "written-beyond", "date": "2026-01-23", "stance": "warn", "quote": "Gemini cli just randomly breaks and starts spamming content trying to reorient itself after a while.", "excerpt": "Has anyone seriously used codex cli? I was using LLMs for code gen usually through the vscode codex extension, Gemini cli and Claude Code cli. The performance of all 3 of them is utter dog shit, Gemini cli just randomly breaks and starts spamming content trying to reorient itself after a while. However, I decided to try codex cli after hearing they rebuilt it from the ground up and used rust(instead of JS, not implyi", "context": "", "use_case": "AI code generation", "switched_from": "", "story": "Unrolling the Codex Agent Loop"}, {"id": 746079872, "item_id": 46629992, "url": "https://news.ycombinator.com/item?id=46629992", "by": "esperent", "date": "2026-01-15", "stance": "warn", "quote": "Then Murphy's law hit me with bug after bug, all of which have been reported many times already going back over year or more.", "excerpt": "I actually defended Gemini CLI when someone said this a few days ago. Then Murphy's law hit me with bug after bug, all of which have been reported many times already going back over year or more. I keep having to totally clear out the ~/.gemini folder and then it works again for a while.", "context": "", "use_case": "CLI coding agent", "switched_from": "", "story": "Ask HN: What is best way to provide continuous context to models?"}, {"id": 781681696, "item_id": 48855106, "url": "https://news.ycombinator.com/item?id=48855106", "by": "ralusek", "date": "2026-07-09", "stance": "warn", "quote": "I don't use Gemini cli because they're so bad at agentic work/tool calling.", "excerpt": "I don't use Gemini cli because they're so bad at agentic work/tool calling. I use their chatbot all the time, though.", "context": "", "use_case": "agentic coding", "switched_from": "", "story": "Introducing Muse Spark 1.1"}]}, {"text": " and some note it is being phased out for Antigravity CLI.", "cites": [{"id": 779929856, "item_id": 48745616, "url": "https://news.ycombinator.com/item?id=48745616", "by": "2muchcoffeeman", "date": "2026-07-01", "stance": "warn", "quote": "I used to use Gemini-cli until they did a Google and cancelled it in favour of anti-gravity.", "excerpt": "The best models are clearly all from US companies. But I think what you\u2019ll see is people making sure the model they use can just be plugged into their workflow. I used to use Gemini-cli until they did a Google and cancelled it in favour of anti-gravity. That was my fault. Fool me the 10th time shame on me. So I picked more open source offerings that I can use with any model. Once the other models are good enough, I j", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5"}, {"id": 775528482, "item_id": 48470530, "url": "https://news.ycombinator.com/item?id=48470530", "by": "kaydub", "date": "2026-06-09", "stance": "neutral", "quote": "gemini cli (which is being phased out for antigravity cli soon)", "excerpt": "I use any of the 3 big frontier models; claude, chatgpt, gemini I only use the cli tools; claude code, codex, gemini cli (which is being phased out for antigravity cli soon) I've generally had the best luck and enjoy working with claude the most, though that may be because that's the model we use at work (from aws bedrock, we pay per token and always have). Gemini and codex have flip flopped on 2nd place, though righ", "context": "", "use_case": "personal projects", "switched_from": "", "story": "Cleaning up after AI rockstar developers"}]}], "praise": [{"label": "Coding capability", "summary": "Some users found it an effective coding agent that makes targeted edits and handled large projects.", "quotes": [{"id": 788680097, "item_id": 49292506, "url": "https://news.ycombinator.com/item?id=49292506", "by": "gazebo2", "date": "2026-08-13", "stance": "recommend", "quote": "found it also an effective coding agent", "excerpt": "I similarly have a weird affinity for Gemini that I can't really articulate. I used Gemini's free chat and found it great for exploring technical topics (and random one-off general walking-around-questions) and appreciated its speed, tone and accuracy. I spent a month playing with Gemini CLI / Antigravity and found it also an effective coding agent, at least for my workflow (entirely in the loop development and revie", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Gemini 3.7 Flash"}, {"id": 736169824, "item_id": 46010614, "url": "https://news.ycombinator.com/item?id=46010614", "by": "keithba", "date": "2025-11-21", "stance": "recommend", "quote": "I used Gemini CLI (once it came, AI Studio before that) with Gemini 2.5 Pro to create one with Rust and LLVM", "excerpt": "LLMs can definitely write programming languages. I used Gemini CLI (once it came, AI Studio before that) with Gemini 2.5 Pro to create one with Rust and LLVM: https://github.com/GoogleCloudPlatform/aether This was completely vibe coded - I never had to edit the code, though it was very interactive. The whole thing tool less than a month of some evenings and weekends. (Note: it\u2019s ugly on purpose, as I\u2019m playing with i", "context": "", "use_case": "vibe coding a language", "switched_from": "AI Studio", "story": "FAWK: LLMs can write a language interpreter"}, {"id": 744672928, "item_id": 46542058, "url": "https://news.ycombinator.com/item?id=46542058", "by": "mattarm", "date": "2026-01-08", "stance": "recommend", "quote": "", "excerpt": "Yeah, the online Gemini app is not good for long lived conversations that build up a body of decisions. The context window gets too large and things drop. What I\u2019ve learned is that once you reach that point you\u2019ve got to break that problem down into smaller pieces that the AI can work productively with. If you\u2019re about to start with Gemini-cli I recommend you look up https://github.com/github/spec-kit. It\u2019s a project", "context": "", "use_case": "AI-assisted document/code editing", "switched_from": "Gemini", "story": "Opus 4.5 is going to change everything"}]}, {"label": "Terminal interface", "summary": "Commenters like its simple text-based terminal interaction, and one calls its terminal UI the nicest.", "quotes": [{"id": 757442993, "item_id": 47340187, "url": "https://news.ycombinator.com/item?id=47340187", "by": "xrd", "date": "2026-03-11", "stance": "recommend", "quote": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage.", "excerpt": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage. Why are they great? Because it is simply text that I use to interact with them. That's really simple and powerful. I don't understand why I can't levitate that simple interface into a web UI inside my phone browser? It feels like this should be as simple as webmux (tmux on the web). But it feels surprisingly elusive. I would really like som", "context": "", "use_case": "terminal coding agent", "switched_from": "", "story": "Karpathy is searching for the Agentic IDE"}, {"id": 735788498, "item_id": 45986781, "url": "https://news.ycombinator.com/item?id=45986781", "by": "freediver", "date": "2025-11-19", "stance": "warn", "quote": "Gemini CLI was never a contender (between log in/activation/rate limits - wth), will say though that Gemini CLI has the nicest terminal UI.", "excerpt": "First time that there is a worthy alternative to Claude Code. Codex Max solved a problem I had Claude Code fail multiple times. Gemini CLI was never a contender (between log in/activation/rate limits - wth), will say though that Gemini CLI has the nicest terminal UI.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Building more with GPT-5.1-Codex-Max"}]}, {"label": "Open and extensible", "summary": "It is open source and actively developed, with extensions, model switching and remote A2A agent support.", "quotes": [{"id": 737029072, "item_id": 46064317, "url": "https://news.ycombinator.com/item?id=46064317", "by": "terrywang", "date": "2025-11-26", "stance": "recommend", "quote": "Gemini CLI is open source and being actively developed, which is cool", "excerpt": "Gemini CLI at this stage isn't good at complex coding tasks (vs. Claude Code, Codex, Cursor CLI, Qoder CLI, etc.). Mostly because of the simple ReAct loop, compounded by relatively weak tool calling capability of the Gemini 2.5 Pro model. > I haven't tried complex coding tasks using Gemini 3.0 Pro Preview yet. I reckon it won't be materially different. Gemini CLI is open source and being actively developed, which is ", "context": "", "use_case": "orchestrating tasks and workflows", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}, {"id": 777488736, "item_id": 48593046, "url": "https://news.ycombinator.com/item?id=48593046", "by": "fhrow4484", "date": "2026-06-18", "stance": "recommend", "quote": "gemini-cli can have remote a2a agents connected to it.", "excerpt": "If you're big in the google ecosystem it makes sense: - gemini-cli can have remote a2a agents connected to it. (See https://github.com/tanaikech/gemini-cli-gas-a2a-subagents and the medium article linked in that repo description.) Haven't tried it but I tried the local use case below. - can expose a custom Google ADK \"agent\" via one-liner as a2a, which you can connect to the gemini-cli the same way.", "context": "Google ecosystem", "use_case": "connecting A2A agents", "switched_from": "", "story": "Ask HN: Is anyone using the A2A protocol yet?"}]}, {"label": "Google ecosystem", "summary": "Users invested in Google's tooling value it alongside NotebookLM, Colab and ADK agents.", "quotes": [{"id": 737937076, "item_id": 46121067, "url": "https://news.ycombinator.com/item?id=46121067", "by": "mark_l_watson", "date": "2025-12-02", "stance": "recommend", "quote": "I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc.", "excerpt": "I used DeepSeek-v3.2 to solve two coding problems by pasting code and directions as one large prompt into a chat interface and it performed very well. VERY WELL! I am still happy to pay Google because of their ecosystem or Gemini app, NotebookLM, Colab, gemini-cli, etc. Google\u2019s moat for me is all the tooling and engineering around the models. That said, my one year Google AI subscription ends in four months and I mi", "context": "retired 'gentleman scientist'", "use_case": "", "switched_from": "", "story": "DeepSeek-v3.2: Pushing the frontier of open large language models [pdf]"}, {"id": 777488736, "item_id": 48593046, "url": "https://news.ycombinator.com/item?id=48593046", "by": "fhrow4484", "date": "2026-06-18", "stance": "recommend", "quote": "gemini-cli can have remote a2a agents connected to it.", "excerpt": "If you're big in the google ecosystem it makes sense: - gemini-cli can have remote a2a agents connected to it. (See https://github.com/tanaikech/gemini-cli-gas-a2a-subagents and the medium article linked in that repo description.) Haven't tried it but I tried the local use case below. - can expose a custom Google ADK \"agent\" via one-liner as a2a, which you can connect to the gemini-cli the same way.", "context": "Google ecosystem", "use_case": "connecting A2A agents", "switched_from": "", "story": "Ask HN: Is anyone using the A2A protocol yet?"}]}, {"label": "Explaining code", "summary": "One user preferred it over Claude Code for concise explanations of a module before diving in.", "quotes": [{"id": 795435617, "item_id": 49714726, "url": "https://news.ycombinator.com/item?id=49714726", "by": "pid0x17", "date": "2026-09-15", "stance": "recommend", "quote": "I use to love using Gemini CLI to have Gemini explain a module for me", "excerpt": "We moved from Gemini to Claude recently, and while I absolutely love Claude's coding capabilities, I often find myself extremely overwhelmed whenever I ask Claude to explain something to me, whereas that was a breeze with Gemini Pro models. It generates pages for what could be explained in 1-2 paragraphs. I use to love using Gemini CLI to have Gemini explain a module for me before I dive into the code and point out s", "context": "team moved from Gemini to Claude", "use_case": "explaining code modules", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}]}], "caveats": [{"label": "Behind competitors", "summary": "Many commenters rate it well below Claude Code, Codex and OpenCode.", "quotes": [{"id": 753690304, "item_id": 47105644, "url": "https://news.ycombinator.com/item?id=47105644", "by": "dudeinhawaii", "date": "2026-02-21", "stance": "warn", "quote": "I find that Gemini CLI is still considerably worse than both Codex and Claude Code.", "excerpt": "After 2 days of giving it a go, I find that Gemini CLI is still considerably worse than both Codex and Claude Code. The model itself also has strange behaviors that seem like it gets randomly replaced with Gemini-3-Flash or something else. I'll explain. Once agentic coding was a bust, I gave it a run as a daily driver for AI assistant. It performed fairly well but then began behaving strangely. It would lose context ", "context": "Google AI Ultra subscriber", "use_case": "agentic coding", "switched_from": "", "story": "Gemini 3.1 Pro Preview"}, {"id": 759506163, "item_id": 47469135, "url": "https://news.ycombinator.com/item?id=47469135", "by": "arthurjean", "date": "2026-03-21", "stance": "warn", "quote": "well above Codex, Gemini CLI or Mistral Vibe in my experience", "excerpt": "I've used both. I stuck with Claude Code, the ergonomics are better and the internals are clearly optimized for Opus which I use daily, you can feel it. That said OpenCode is still a very good alternative, well above Codex, Gemini CLI or Mistral Vibe in my experience.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "OpenCode \u2013 The open source AI coding agent"}, {"id": 744514739, "item_id": 46532171, "url": "https://news.ycombinator.com/item?id=46532171", "by": "vedmakk", "date": "2026-01-07", "stance": "warn", "quote": "Gemini CLI is a pain to use - after getting used to the smoothness of Claude\nCode.", "excerpt": "I don't get the Gemini 3 hype... yes it's their first usable model, but its not even close to what Opus 4.5 and GPT 5.2 can do. Maybe on Benchmarks... but I'm forced to use Gemini at work everyday, while I use Opus 4.5 / GPT 5.2 privately every day... and Gemini is just lacking so much wit, creativity and multi-step problem solving skills compared to Opus. Not to mention that Gemini CLI is a pain to use - after getti", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Google Got Its Groove Back and Edged Ahead of OpenAI"}]}, {"label": "Bugs and crashes", "summary": "Users report crashes, infinite loops, lost chat history and long-standing unfixed bugs.", "quotes": [{"id": 786995440, "item_id": 49187215, "url": "https://news.ycombinator.com/item?id=49187215", "by": "pixel_popping", "date": "2026-08-05", "stance": "warn", "quote": "Gemini CLI was utter shit and always had delayed model release on top of it.", "excerpt": "Gemini CLI was utter shit and always had delayed model release on top of it. I'm not kidding when 1 hour automated vibe-coding led to an agent 30% cheaper and faster on 1:1 tasks with the SAME Gemini models (re-ran on 50 past tasks), genuinely shameful that Google even released this, it was so buggy, trapped in infinite loops as well and so-on. Antigravity is a whole different product, the only way I use it currently", "context": "", "use_case": "agentic coding CLI", "switched_from": "", "story": "Antigravity 2.0"}, {"id": 746079872, "item_id": 46629992, "url": "https://news.ycombinator.com/item?id=46629992", "by": "esperent", "date": "2026-01-15", "stance": "warn", "quote": "Then Murphy's law hit me with bug after bug, all of which have been reported many times already going back over year or more.", "excerpt": "I actually defended Gemini CLI when someone said this a few days ago. Then Murphy's law hit me with bug after bug, all of which have been reported many times already going back over year or more. I keep having to totally clear out the ~/.gemini folder and then it works again for a while.", "context": "", "use_case": "CLI coding agent", "switched_from": "", "story": "Ask HN: What is best way to provide continuous context to models?"}, {"id": 734770721, "item_id": 45923170, "url": "https://news.ycombinator.com/item?id=45923170", "by": "ilyakaminsky", "date": "2025-11-13", "stance": "neutral", "quote": "I use Gemini CLI on a daily basis. It used to crash often", "excerpt": "I use Gemini CLI on a daily basis. It used to crash often and I'd lose the chat history. I found this tool called ai-cli-log [1] and it does something similar out of the box. I don't run Gemini CLI without it. [1] https://github.com/alingse/ai-cli-log", "context": "", "use_case": "AI coding in terminal", "switched_from": "", "story": "Nano Banana can be prompt engineered for nuanced AI image generation"}]}, {"label": "Weak agentic work", "summary": "Commenters say it is poor at tool calling, codebase navigation and complex agentic tasks.", "quotes": [{"id": 781681696, "item_id": 48855106, "url": "https://news.ycombinator.com/item?id=48855106", "by": "ralusek", "date": "2026-07-09", "stance": "warn", "quote": "I don't use Gemini cli because they're so bad at agentic work/tool calling.", "excerpt": "I don't use Gemini cli because they're so bad at agentic work/tool calling. I use their chatbot all the time, though.", "context": "", "use_case": "agentic coding", "switched_from": "", "story": "Introducing Muse Spark 1.1"}, {"id": 746746737, "item_id": 46671671, "url": "https://news.ycombinator.com/item?id=46671671", "by": "basher", "date": "2026-01-18", "stance": "warn", "quote": "I tried to use the official `gemini-cli` to avoid the API costs. It is unusable for agentic work.", "excerpt": "I\u2019ve been paying *$125/mo* (promo pricing for the $250/mo tier) for Google\u2019s top-tier \"AI Ultra\" subscription for a month. I signed up because I\u2019m a power user\u2014I manage large repos, build my own MCP tools, and live in the terminal. I assumed \"Ultra\" meant \"Pro Developer.\" I was wrong. I downgraded to the $20/mo plan today, and here is why other developers should probably avoid the top tier. 1. The \"Double Pay\" Parado", "context": "power user managing large repos and building MCP tools", "use_case": "agentic coding in terminal", "switched_from": "", "story": "Google's \"AI Ultra\" Is a Storage Plan Disguised as a Dev Tool"}, {"id": 737045889, "item_id": 46065368, "url": "https://news.ycombinator.com/item?id=46065368", "by": "MarkMarine", "date": "2025-11-26", "stance": "warn", "quote": "I tried Gemini cli for about 5 seconds and was so frustrated, it\u2019s so stupid at navigation in the codebase", "excerpt": "I love the model, hate the tool. I\u2019ve taken complex stuff and given it to Gemini 3 and been impressed, but Anthropic has the killer app with Claude Code. The interplay of sonnet (a decent model) and the tools and workflow they\u2019ve got with Claude code around it supercharge the outcome. I tried Gemini cli for about 5 seconds and was so frustrated, it\u2019s so stupid at navigation in the codebase it takes 10x as long to do ", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}]}, {"label": "Loses context", "summary": "It reportedly forgets prior prompts, ignores GEMINI.md files and hallucinates.", "quotes": [{"id": 754707952, "item_id": 47169247, "url": "https://news.ycombinator.com/item?id=47169247", "by": "anonzzzies", "date": "2026-02-26", "stance": "warn", "quote": "Gemini cli (paid) often forgets by the next prompt whatever the previous was about", "excerpt": "Gemini is supposed to have this huge context; Gemini cli (paid) often forgets by the next prompt whatever the previous was about and starts doing something completely different , often switching natural or programming language. I use codex and with 5.3 it is better but not there compared to cc for us anyway; it just goes looking for stuff, draws the most bizarre conclusions and ends up lost quite often doing the wron", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Why Developers Keep Choosing Claude over Every Other AI"}, {"id": 753690304, "item_id": 47105644, "url": "https://news.ycombinator.com/item?id=47105644", "by": "dudeinhawaii", "date": "2026-02-21", "stance": "warn", "quote": "I find that Gemini CLI is still considerably worse than both Codex and Claude Code.", "excerpt": "After 2 days of giving it a go, I find that Gemini CLI is still considerably worse than both Codex and Claude Code. The model itself also has strange behaviors that seem like it gets randomly replaced with Gemini-3-Flash or something else. I'll explain. Once agentic coding was a bust, I gave it a run as a daily driver for AI assistant. It performed fairly well but then began behaving strangely. It would lose context ", "context": "Google AI Ultra subscriber", "use_case": "agentic coding", "switched_from": "", "story": "Gemini 3.1 Pro Preview"}, {"id": 737061904, "item_id": 46066369, "url": "https://news.ycombinator.com/item?id=46066369", "by": "INTPenis", "date": "2025-11-26", "stance": "recommend", "quote": "gemini cli gave me my first real fun experience using AI", "excerpt": "I've been using Gemini CLI for months now, mainly because we have a free subscription for it through work. Tip 1, it consistently ignores my GEMINI.md file, both global and local. Even though it's always saying that \"1 GEMINI.md file is being used\", probably because the file exists in the right path. Tip 12, had no idea you could do this, seems like a great tip to me. Tip 16 was great, thanks. I've been restarting it", "context": "has free work subscription", "use_case": "troubleshooting k8s cluster", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}]}, {"label": "Tokens and limits", "summary": "Users cite heavy token consumption, rate-limit errors, slowness and a painful login and setup.", "quotes": [{"id": 750760192, "item_id": 46922512, "url": "https://news.ycombinator.com/item?id=46922512", "by": "throwa356262", "date": "2026-02-07", "stance": "warn", "quote": "I am reluctant to fire up Gemini CLI due to its insane appetite for tokens.", "excerpt": "I need to give this a try, but nowadays I am reluctant to fire up Gemini CLI due to its insane appetite for tokens. It doesnt matter if your LLM in/out tokens are a bit cheaper than competitors when you use 3x of them on every prompt. Maybe Google should focus on addressing that first?", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Introducing the Developer Knowledge API and MCP Server"}, {"id": 755190304, "item_id": 47199394, "url": "https://news.ycombinator.com/item?id=47199394", "by": "hsaliak", "date": "2026-02-28", "stance": "warn", "quote": "the 429s is the first thing they need to fix.", "excerpt": "It takes your query, computes the complexity of the request, and tries to route it to the appropriate model. There is a /manual command i think, to pick the right model. They mask the 429s well in Gemini-Cli - if an endpoint is rate limited, they try another, or route to another model, etc to keep service availability up. Your experience on the 429s is consistent with mine - the 429s is the first thing they need to f", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Addressing Antigravity Bans and Reinstating Access"}, {"id": 744604512, "item_id": 46537782, "url": "https://news.ycombinator.com/item?id=46537782", "by": "mythz", "date": "2026-01-07", "stance": "warn", "quote": "Gemini CLI is too slow to be useful", "excerpt": "Gemini CLI is too slow to be useful, kind of surprised it was even offered and marketed given how painful it is to use. I thought it'd have to be damaging to the Gemini brand to get people to try it out, suffer painful UX then immediately stop using it. (Using it from Australia may also contribute to its slow perf) Antigravity was also painful to use at launch where more queries failed then succeeded, however they've", "context": "using from Australia", "use_case": "AI coding", "switched_from": "", "story": "Google Got Its Groove Back and Edged Ahead of OpenAI"}]}, {"label": "Being discontinued", "summary": "Commenters say Google is phasing it out in favor of Antigravity CLI, raising longevity concerns.", "quotes": [{"id": 779929856, "item_id": 48745616, "url": "https://news.ycombinator.com/item?id=48745616", "by": "2muchcoffeeman", "date": "2026-07-01", "stance": "warn", "quote": "I used to use Gemini-cli until they did a Google and cancelled it in favour of anti-gravity.", "excerpt": "The best models are clearly all from US companies. But I think what you\u2019ll see is people making sure the model they use can just be plugged into their workflow. I used to use Gemini-cli until they did a Google and cancelled it in favour of anti-gravity. That was my fault. Fool me the 10th time shame on me. So I picked more open source offerings that I can use with any model. Once the other models are good enough, I j", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5"}, {"id": 775528482, "item_id": 48470530, "url": "https://news.ycombinator.com/item?id=48470530", "by": "kaydub", "date": "2026-06-09", "stance": "neutral", "quote": "gemini cli (which is being phased out for antigravity cli soon)", "excerpt": "I use any of the 3 big frontier models; claude, chatgpt, gemini I only use the cli tools; claude code, codex, gemini cli (which is being phased out for antigravity cli soon) I've generally had the best luck and enjoy working with claude the most, though that may be because that's the model we use at work (from aws bedrock, we pay per token and always have). Gemini and codex have flip flopped on 2nd place, though righ", "context": "", "use_case": "personal projects", "switched_from": "", "story": "Cleaning up after AI rockstar developers"}, {"id": 796576289, "item_id": 49786018, "url": "https://news.ycombinator.com/item?id=49786018", "by": "fg137", "date": "2026-09-21", "stance": "warn", "quote": "Especially if you look at what happened to Android, Gemini CLI etc.", "excerpt": "Yeah, anyone should think of this as an experimental side project that may be forgotten in 15min and make careful decisions about using them in production environments. We have used a few of Google's (smaller) open source projects, and in the last 2-3 years most of them are getting fewer updates if any updates at all. Some became very bad tech debt and we had to spend a lot of time migrating them. Of course, that is ", "context": "", "use_case": "", "switched_from": "", "story": "AX \u2013 Google\u2019s Open Agentic Orchestrator"}]}], "who": [{"by": "pid0x17", "context": "team moved from Gemini to Claude", "url": "https://news.ycombinator.com/item?id=49714726", "date": "2026-09-15"}, {"by": "fhrow4484", "context": "Google ecosystem", "url": "https://news.ycombinator.com/item?id=48593046", "date": "2026-06-18"}, {"by": "lywald", "context": "solo creator", "url": "https://news.ycombinator.com/item?id=46757306", "date": "2026-01-25"}, {"by": "game_the0ry", "context": "pays for the $20 plan", "url": "https://news.ycombinator.com/item?id=46755913", "date": "2026-01-25"}, {"by": "mark_l_watson", "context": "40 years of Emacs experience", "url": "https://news.ycombinator.com/item?id=46499746", "date": "2026-01-05"}, {"by": "INTPenis", "context": "has free work subscription", "url": "https://news.ycombinator.com/item?id=46066369", "date": "2025-11-26"}], "tags": ["for terminal users", "for Google ecosystem users", "for multi-agent workflows", "for explaining code"], "use_cases": ["ai coding", "coding agent", "coding", "agentic coding", "cli coding agent", "shared agent instructions"], "compared": [{"slug": "claude-code", "name": "Claude Code", "co": 37, "switched_from": 0, "switched_to": 2}, {"slug": "codex", "name": "Codex", "co": 26, "switched_from": 0, "switched_to": 1}, {"slug": "gemini", "name": "Gemini", "co": 10, "switched_from": 1, "switched_to": 0}, {"slug": "antigravity", "name": "Antigravity", "co": 7, "switched_from": 0, "switched_to": 2}, {"slug": "opencode", "name": "OpenCode", "co": 7, "switched_from": 0, "switched_to": 0}, {"slug": "claude", "name": "Claude", "co": 4, "switched_from": 1, "switched_to": 0}, {"slug": "gemini-pro", "name": "Gemini Pro", "co": 5, "switched_from": 0, "switched_to": 0}, {"slug": "gpt-5", "name": "GPT-5", "co": 5, "switched_from": 0, "switched_to": 0}], "makers": [], "threads": [{"url": "https://news.ycombinator.com/item?id=46060508", "title": "Gemini CLI tips and tricks for agentic coding", "mentions": 4}, {"url": "https://news.ycombinator.com/item?id=45917875", "title": "Nano Banana can be prompt engineered for nuanced AI image generation", "mentions": 3}, {"url": "https://news.ycombinator.com/item?id=46626639", "title": "Ask HN: What is best way to provide continuous context to models?", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=46528389", "title": "Google Got Its Groove Back and Edged Ahead of OpenAI", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=46498735", "title": "I switched from VSCode to Zed", "mentions": 2}, {"url": "https://news.ycombinator.com/item?id=49780797", "title": "AX \u2013 Google\u2019s Open Agentic Orchestrator", "mentions": 1}], "sentiment_by_month": {"2025-11": {"neutral": 2, "recommend": 4, "warn": 3}, "2025-12": {"recommend": 2, "warn": 4, "neutral": 1}, "2026-01": {"neutral": 3, "warn": 7, "recommend": 3}, "2026-02": {"warn": 5, "neutral": 2}, "2026-03": {"neutral": 4, "recommend": 1, "warn": 1}, "2026-04": {"neutral": 1}, "2026-05": {"warn": 1}, "2026-06": {"neutral": 2, "recommend": 1, "warn": 1}, "2026-07": {"warn": 2}, "2026-08": {"warn": 1, "recommend": 1}, "2026-09": {"neutral": 2, "recommend": 1, "warn": 1}, "2026-10": {}}, "quotes": [{"id": 796576289, "item_id": 49786018, "url": "https://news.ycombinator.com/item?id=49786018", "by": "fg137", "date": "2026-09-21", "stance": "warn", "quote": "Especially if you look at what happened to Android, Gemini CLI etc.", "excerpt": "Yeah, anyone should think of this as an experimental side project that may be forgotten in 15min and make careful decisions about using them in production environments. We have used a few of Google's (smaller) open source projects, and in the last 2-3 years most of them are getting fewer updates if any updates at all. Some became very bad tech debt and we had to spend a lot of time migrating them. Of course, that is ", "context": "", "use_case": "", "switched_from": "", "story": "AX \u2013 Google\u2019s Open Agentic Orchestrator"}, {"id": 795763300, "item_id": 49735206, "url": "https://news.ycombinator.com/item?id=49735206", "by": "chrismarlow9", "date": "2026-09-17", "stance": "neutral", "quote": "Symlink the global MD files to a single file I control that is versioned.", "excerpt": "Symlink the global MD files to a single file I control that is versioned. Define my own general spec and workflow terminology using markdown files in this universal MD. ``` ~/.config/opencode/AGENTS.md ~/.claude/CLAUDE.md ~/.codex/AGENTS.md ~/.copilot/copilot-instructions.md ~/.gemini/GEMINI.md ``` Define it as a graph and iterate. I use more tokens, but I can also use more tools without disruption. Delegating markdo", "context": "", "use_case": "shared agent instructions", "switched_from": "", "story": "OpenSpec \u2013 A lightweight and configurable AI spec framework"}, {"id": 795435617, "item_id": 49714726, "url": "https://news.ycombinator.com/item?id=49714726", "by": "pid0x17", "date": "2026-09-15", "stance": "recommend", "quote": "I use to love using Gemini CLI to have Gemini explain a module for me", "excerpt": "We moved from Gemini to Claude recently, and while I absolutely love Claude's coding capabilities, I often find myself extremely overwhelmed whenever I ask Claude to explain something to me, whereas that was a breeze with Gemini Pro models. It generates pages for what could be explained in 1-2 paragraphs. I use to love using Gemini CLI to have Gemini explain a module for me before I dive into the code and point out s", "context": "team moved from Gemini to Claude", "use_case": "explaining code modules", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}, {"id": 794619955, "item_id": 49663747, "url": "https://news.ycombinator.com/item?id=49663747", "by": "phudyka", "date": "2026-09-11", "stance": "neutral", "quote": "I have also used Codex, Cursor and Gemini CLI.", "excerpt": "Location: France (CET) Remote: Yes, fully remote only (occasional travel/offsites fine) Willing to relocate: No Technologies: Flutter/Dart, Supabase (PostgreSQL, RLS, Deno Edge Functions), TypeScript, React, Next.js, Django, Angular, Docker, GitHub Actions, n8n R\u00e9sum\u00e9/CV: on request. linkedin.com/in/paul-hudyka, github.com/phudyka, phudyka.github.io Email: phudyka.dev@gmail.com Web and mobile developer. I ship produc", "context": "", "use_case": "", "switched_from": "", "story": "Ask HN: Who wants to be hired? (September 2026)"}, {"id": 788680097, "item_id": 49292506, "url": "https://news.ycombinator.com/item?id=49292506", "by": "gazebo2", "date": "2026-08-13", "stance": "recommend", "quote": "found it also an effective coding agent", "excerpt": "I similarly have a weird affinity for Gemini that I can't really articulate. I used Gemini's free chat and found it great for exploring technical topics (and random one-off general walking-around-questions) and appreciated its speed, tone and accuracy. I spent a month playing with Gemini CLI / Antigravity and found it also an effective coding agent, at least for my workflow (entirely in the loop development and revie", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Gemini 3.7 Flash"}, {"id": 786995440, "item_id": 49187215, "url": "https://news.ycombinator.com/item?id=49187215", "by": "pixel_popping", "date": "2026-08-05", "stance": "warn", "quote": "Gemini CLI was utter shit and always had delayed model release on top of it.", "excerpt": "Gemini CLI was utter shit and always had delayed model release on top of it. I'm not kidding when 1 hour automated vibe-coding led to an agent 30% cheaper and faster on 1:1 tasks with the SAME Gemini models (re-ran on 50 past tasks), genuinely shameful that Google even released this, it was so buggy, trapped in infinite loops as well and so-on. Antigravity is a whole different product, the only way I use it currently", "context": "", "use_case": "agentic coding CLI", "switched_from": "", "story": "Antigravity 2.0"}, {"id": 781681696, "item_id": 48855106, "url": "https://news.ycombinator.com/item?id=48855106", "by": "ralusek", "date": "2026-07-09", "stance": "warn", "quote": "I don't use Gemini cli because they're so bad at agentic work/tool calling.", "excerpt": "I don't use Gemini cli because they're so bad at agentic work/tool calling. I use their chatbot all the time, though.", "context": "", "use_case": "agentic coding", "switched_from": "", "story": "Introducing Muse Spark 1.1"}, {"id": 779929856, "item_id": 48745616, "url": "https://news.ycombinator.com/item?id=48745616", "by": "2muchcoffeeman", "date": "2026-07-01", "stance": "warn", "quote": "I used to use Gemini-cli until they did a Google and cancelled it in favour of anti-gravity.", "excerpt": "The best models are clearly all from US companies. But I think what you\u2019ll see is people making sure the model they use can just be plugged into their workflow. I used to use Gemini-cli until they did a Google and cancelled it in favour of anti-gravity. That was my fault. Fool me the 10th time shame on me. So I picked more open source offerings that I can use with any model. Once the other models are good enough, I j", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5"}, {"id": 779002161, "item_id": 48687635, "url": "https://news.ycombinator.com/item?id=48687635", "by": "sgathrid", "date": "2026-06-26", "stance": "neutral", "quote": "using Gemini-CLI w/ gemini-3.1-pro-preview", "excerpt": "Refactoring code with Topos cleans up your program structure. Agents optimize for simpler ASTs and lower coupling scores via MCP. We ran a controlled experiment: implement 4 features into two identically-functioning repos -- one that had been refactored using Topos and one that has not been -- using Gemini-CLI w/ gemini-3.1-pro-preview. Findings: - 32.6% fewer tokens to add features to the Topos-optimized repo - 22.9", "context": "", "use_case": "running the experiment", "switched_from": "", "story": "Module decomposition cut agent token use 32% on follow-up feature additions"}, {"id": 777721985, "item_id": 48607624, "url": "https://news.ycombinator.com/item?id=48607624", "by": "bjackman", "date": "2026-06-20", "stance": "warn", "quote": "it's a huge step up from Gemini CLI (like, for example, it actually fucking works)", "excerpt": "This is weird coz as a user of both Gemini and Claude I have the opposite feeling. Antigravity CLI is quite decent, it's a huge step up from Gemini CLI (like, for example, it actually fucking works) and has some genuine advantages over Claude Code. Does Codex have something over both of them? I haven't tried it. But the model just fucking sucks. Before I switched to Claude for personal stuff a few weeks ago, I was li", "context": "", "use_case": "AI coding", "switched_from": "", "story": "John Jumper leaves Google to join Anthropic"}, {"id": 777488736, "item_id": 48593046, "url": "https://news.ycombinator.com/item?id=48593046", "by": "fhrow4484", "date": "2026-06-18", "stance": "recommend", "quote": "gemini-cli can have remote a2a agents connected to it.", "excerpt": "If you're big in the google ecosystem it makes sense: - gemini-cli can have remote a2a agents connected to it. (See https://github.com/tanaikech/gemini-cli-gas-a2a-subagents and the medium article linked in that repo description.) Haven't tried it but I tried the local use case below. - can expose a custom Google ADK \"agent\" via one-liner as a2a, which you can connect to the gemini-cli the same way.", "context": "Google ecosystem", "use_case": "connecting A2A agents", "switched_from": "", "story": "Ask HN: Is anyone using the A2A protocol yet?"}, {"id": 775528482, "item_id": 48470530, "url": "https://news.ycombinator.com/item?id=48470530", "by": "kaydub", "date": "2026-06-09", "stance": "neutral", "quote": "gemini cli (which is being phased out for antigravity cli soon)", "excerpt": "I use any of the 3 big frontier models; claude, chatgpt, gemini I only use the cli tools; claude code, codex, gemini cli (which is being phased out for antigravity cli soon) I've generally had the best luck and enjoy working with claude the most, though that may be because that's the model we use at work (from aws bedrock, we pay per token and always have). Gemini and codex have flip flopped on 2nd place, though righ", "context": "", "use_case": "personal projects", "switched_from": "", "story": "Cleaning up after AI rockstar developers"}, {"id": 771414368, "item_id": 48213398, "url": "https://news.ycombinator.com/item?id=48213398", "by": "almog", "date": "2026-05-20", "stance": "warn", "quote": "It felt far behind both CC and Codex", "excerpt": "Earlier this month I switched from Claude Code to Codex and wanted to try Gemini CLI as well. It felt far behind both CC and Codex but I wanted to give it another chance with the new Antigravity CLI. What can I say, it did surprise me and not in a good way in but two short sessions that included just two prompts (trying to reverse engineer some earbuds OTA firmware) using Gemini Flash 3.5, I managed to finish my week", "context": "on Google AI Pro subscription", "use_case": "AI coding CLI", "switched_from": "", "story": "Gemini CLI will stop working from June 18, 2026"}, {"id": 767262290, "item_id": 47953893, "url": "https://news.ycombinator.com/item?id=47953893", "by": "christoff12", "date": "2026-04-29", "stance": "neutral", "quote": "running Codex + Gemini CLI in the terminal panel", "excerpt": "I like Zed for the same reason I liked Atom: it's very light. These past few months my workflow has basically consisted of me running Codex + Gemini CLI in the terminal panel and hopping through files in another. Easy-peasy. It's a nice departure from the visual overstimulation I get while in VS Code (for which I have to take some blame as I need to remove some installed plugins). There are a couple of features shown", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Zed is 1.0"}, {"id": 761491201, "item_id": 47593200, "url": "https://news.ycombinator.com/item?id=47593200", "by": "valentinconan", "date": "2026-03-31", "stance": "neutral", "quote": "I used to use Gemini CLI and was pretty happy with it", "excerpt": "I used to use Gemini CLI and was pretty happy with it\u2026 until I started using Claude Code. I think it\u2019s, for now, the most advanced AI tool available right now. You can manage skills, the MCP, plugins, agents, hooks\u2026 It\u2019s a truly comprehensive tool, but its main drawback is the token consumption of the Anthropic model, especially in recent days. There are solutions to reduce input tokens, for example this GitHub proje", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Ask HN: Gemini CLI vs. Claude Code"}, {"id": 759506163, "item_id": 47469135, "url": "https://news.ycombinator.com/item?id=47469135", "by": "arthurjean", "date": "2026-03-21", "stance": "warn", "quote": "well above Codex, Gemini CLI or Mistral Vibe in my experience", "excerpt": "I've used both. I stuck with Claude Code, the ergonomics are better and the internals are clearly optimized for Opus which I use daily, you can feel it. That said OpenCode is still a very good alternative, well above Codex, Gemini CLI or Mistral Vibe in my experience.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "OpenCode \u2013 The open source AI coding agent"}, {"id": 758010626, "item_id": 47375664, "url": "https://news.ycombinator.com/item?id=47375664", "by": "mark_l_watson", "date": "2026-03-14", "stance": "neutral", "quote": "several times a week have sessions with gemini-cli and Claude Code when I want an AI coding assistant", "excerpt": "Nice writeup! The list of Emacs packages to use LLMs is the roadmap for getting started with AI Emacs integrations and I also liked his list of AI dangers at the end. I have used Emacs for over 40 years but I found myself getting sucked into the black hole of VSCode as discussed in the article. When I bought a new MacBook a few months ago, I only installed Emacs, no VSCode, and so far that is working out for me. Also", "context": "", "use_case": "AI coding assistant", "switched_from": "", "story": "Emacs and Vim in the Age of AI"}, {"id": 757442993, "item_id": 47340187, "url": "https://news.ycombinator.com/item?id=47340187", "by": "xrd", "date": "2026-03-11", "stance": "recommend", "quote": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage.", "excerpt": "I think Claude Code and Gemini CLI are pretty great as is for terminal usage. Why are they great? Because it is simply text that I use to interact with them. That's really simple and powerful. I don't understand why I can't levitate that simple interface into a web UI inside my phone browser? It feels like this should be as simple as webmux (tmux on the web). But it feels surprisingly elusive. I would really like som", "context": "", "use_case": "terminal coding agent", "switched_from": "", "story": "Karpathy is searching for the Agentic IDE"}, {"id": 757095521, "item_id": 47318470, "url": "https://news.ycombinator.com/item?id=47318470", "by": "radio879", "date": "2026-03-09", "stance": "neutral", "quote": "5x gemini cli in pro 3.1 then 5x flash 3.1", "excerpt": "... also, instead of just using all this stuff out there, like all the OpenClaw stuff, memory systems, what I am doing so that I understand it is just create my own from scratch I have not even read about how that one works, I spin up 10x copilot -p \"idea\", in the new automate mode they just added, in Docker container, 5x gemini cli in pro 3.1 then 5x flash 3.1, gpt 5.4 in there somewhere along with 5.3 codex, o3, 04", "context": "", "use_case": "parallel idea generation in best-of-N", "switched_from": "", "story": "Ask HN: Read\u2011only LLM tool for email triage and knowledge extraction?"}, {"id": 755970853, "item_id": 47248178, "url": "https://news.ycombinator.com/item?id=47248178", "by": "woxpas", "date": "2026-03-04", "stance": "neutral", "quote": "sometimes even Gemini CLI for low level scripting", "excerpt": "I built Woxpas because I got tired of re-explaining myself every time I switched between LLMs. I use Claude for writing, ChatGPT for quick lookups, Cursor or VSC with claude code for coding, and sometimes even Gemini CLI for low level scripting. Each one has its own memory \u2014 none of them talk to each other. Every new session is a cold start. Woxpas is a single memory vault that sits between you and whatever AI you us", "context": "", "use_case": "low-level scripting", "switched_from": "", "story": null}, {"id": 755190304, "item_id": 47199394, "url": "https://news.ycombinator.com/item?id=47199394", "by": "hsaliak", "date": "2026-02-28", "stance": "warn", "quote": "the 429s is the first thing they need to fix.", "excerpt": "It takes your query, computes the complexity of the request, and tries to route it to the appropriate model. There is a /manual command i think, to pick the right model. They mask the 429s well in Gemini-Cli - if an endpoint is rate limited, they try another, or route to another model, etc to keep service availability up. Your experience on the 429s is consistent with mine - the 429s is the first thing they need to f", "context": "", "use_case": "AI coding agent", "switched_from": "", "story": "Addressing Antigravity Bans and Reinstating Access"}, {"id": 754707952, "item_id": 47169247, "url": "https://news.ycombinator.com/item?id=47169247", "by": "anonzzzies", "date": "2026-02-26", "stance": "warn", "quote": "Gemini cli (paid) often forgets by the next prompt whatever the previous was about", "excerpt": "Gemini is supposed to have this huge context; Gemini cli (paid) often forgets by the next prompt whatever the previous was about and starts doing something completely different , often switching natural or programming language. I use codex and with 5.3 it is better but not there compared to cc for us anyway; it just goes looking for stuff, draws the most bizarre conclusions and ends up lost quite often doing the wron", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Why Developers Keep Choosing Claude over Every Other AI"}, {"id": 754626754, "item_id": 47164172, "url": "https://news.ycombinator.com/item?id=47164172", "by": "gck1", "date": "2026-02-26", "stance": "warn", "quote": "Every proprietary harness is just proprietary junk without ability to extend it without polluting context. This includes claude-code, gemini-cli, codex etc.", "excerpt": "Every proprietary harness is just proprietary junk without ability to extend it without polluting context. This includes claude-code, gemini-cli, codex etc. They have tools which hardcode the behavior that is impossible to modify, they add tools you may not need that pollute context, they inject an entire textbook's worth of words into the system prompt which pollutes context, they provide zero observability into wha", "context": "", "use_case": "coding agent harness", "switched_from": "", "story": "How Will OpenAI Compete?"}, {"id": 753690304, "item_id": 47105644, "url": "https://news.ycombinator.com/item?id=47105644", "by": "dudeinhawaii", "date": "2026-02-21", "stance": "warn", "quote": "I find that Gemini CLI is still considerably worse than both Codex and Claude Code.", "excerpt": "After 2 days of giving it a go, I find that Gemini CLI is still considerably worse than both Codex and Claude Code. The model itself also has strange behaviors that seem like it gets randomly replaced with Gemini-3-Flash or something else. I'll explain. Once agentic coding was a bust, I gave it a run as a daily driver for AI assistant. It performed fairly well but then began behaving strangely. It would lose context ", "context": "Google AI Ultra subscriber", "use_case": "agentic coding", "switched_from": "", "story": "Gemini 3.1 Pro Preview"}, {"id": 753652210, "item_id": 47103263, "url": "https://news.ycombinator.com/item?id=47103263", "by": "Bumblebiber", "date": "2026-02-21", "stance": "neutral", "quote": "I switch between Claude Code, Gemini CLI, and OpenCode depending on the task.", "excerpt": "Author here, happy to answer questions. Some background: I run a multi-agent AI system (orchestrator + specialized agents) across multiple machines. Two things kept biting me: 1. *Context dilution:* In long sessions, earlier context gets compressed or dropped. The agent \"forgets\" decisions made hours ago \u2014 not because the session ended, but because the context window silently pushed them out. 2. *Vendor/machine lock-", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Show HN: Hmem \u2013 Persistent hierarchical memory for AI coding agents (MCP)"}, {"id": 753297745, "item_id": 47081109, "url": "https://news.ycombinator.com/item?id=47081109", "by": "operatingthetan", "date": "2026-02-19", "stance": "neutral", "quote": "I've been doing something similar with the Claude or Gemini CLI tools operating in an Obsidian vault.", "excerpt": "I've been doing something similar with the Claude or Gemini CLI tools operating in an Obsidian vault. I use them side by side and can work on the files at the same time.", "context": "", "use_case": "editing notes in vault", "switched_from": "", "story": "Cothought: Claude as text editor, thinking journal"}, {"id": 750760192, "item_id": 46922512, "url": "https://news.ycombinator.com/item?id=46922512", "by": "throwa356262", "date": "2026-02-07", "stance": "warn", "quote": "I am reluctant to fire up Gemini CLI due to its insane appetite for tokens.", "excerpt": "I need to give this a try, but nowadays I am reluctant to fire up Gemini CLI due to its insane appetite for tokens. It doesnt matter if your LLM in/out tokens are a bit cheaper than competitors when you use 3x of them on every prompt. Maybe Google should focus on addressing that first?", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Introducing the Developer Knowledge API and MCP Server"}, {"id": 749323906, "item_id": 46832744, "url": "https://news.ycombinator.com/item?id=46832744", "by": "mag123c", "date": "2026-01-30", "stance": "neutral", "quote": "track my daily token usage across Claude Code, Codex, and Gemini CLI", "excerpt": "I built a local CLI tool to track my daily token usage across Claude Code, Codex, and Gemini CLI. Parsing the session logs directly (JSONL/JSON) with simd-json gives me exact numbers without relying on any external API. Knowing my actual spend per day changed how I use these tools \u2014 I was burning 3x more on cache misses than I realized.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Ask HN: How do you budget for token based AI APIs?"}, {"id": 748116898, "item_id": 46757306, "url": "https://news.ycombinator.com/item?id=46757306", "by": "lywald", "date": "2026-01-25", "stance": "recommend", "quote": "Claude was down so I used Gemini CLI and it turns out it is quite capable also.", "excerpt": "Hey HN, Pierre here. I\u2019m the solo creator of EbookForge. I know the skepticism around AI-generated content (and \"slop\") is high right now. I share it. But I built this tool because I found that for self-education and structuring ideas, LLMs are actually incredible. I used EbookForge itself to generate a massive, 150-page catalog containing 1,000+ book concepts. This goes beyond concept: we can then recursively genera", "context": "solo creator", "use_case": "vibe-coding", "switched_from": "Claude", "story": "Show HN: I used my book generator to generate a catalog of books it can generate"}, {"id": 748094608, "item_id": 46755913, "url": "https://news.ycombinator.com/item?id=46755913", "by": "game_the0ry", "date": "2026-01-25", "stance": "recommend", "quote": "For the CLI -- no, I have not noticed and hallucinations.", "excerpt": "Depends on the context -- do you mean Gemini CLI for coding or Gemini the chat web app? For the CLI -- no, I have not noticed and hallucinations. For the chat app -- no, and the more I use it, the more it feels like its getting better with time. (I pay for the $20 plan btw)", "context": "pays for the $20 plan", "use_case": "coding", "switched_from": "", "story": "Ask HN: Is Gemini Getting Worse?"}, {"id": 747812609, "item_id": 46738288, "url": "https://news.ycombinator.com/item?id=46738288", "by": "written-beyond", "date": "2026-01-23", "stance": "warn", "quote": "Gemini cli just randomly breaks and starts spamming content trying to reorient itself after a while.", "excerpt": "Has anyone seriously used codex cli? I was using LLMs for code gen usually through the vscode codex extension, Gemini cli and Claude Code cli. The performance of all 3 of them is utter dog shit, Gemini cli just randomly breaks and starts spamming content trying to reorient itself after a while. However, I decided to try codex cli after hearing they rebuilt it from the ground up and used rust(instead of JS, not implyi", "context": "", "use_case": "AI code generation", "switched_from": "", "story": "Unrolling the Codex Agent Loop"}, {"id": 747111297, "item_id": 46694456, "url": "https://news.ycombinator.com/item?id=46694456", "by": "sixtyj", "date": "2026-01-20", "stance": "warn", "quote": "Gemini CLI tells me that it doesn't have this and that - I have tried pushing Gemini to deliver production-quality code. No chance.", "excerpt": "This is really odd. They are incomparable in terms of the scope that these companies have. I have used all major tools: OpenAI (chat, api), Google Gemini (ai studio, api, cli, antigravity) and Claude (chat, code and api). Mostly for coding issues to solve. Claude Code gives usable results almost instantly for small scripts and it can live. Gemini CLI tells me that it doesn't have this and that - I have tried pushing ", "context": "", "use_case": "coding", "switched_from": "", "story": "Claude Code is the ChatGPT moment repeated and awful news for software stocks"}, {"id": 746746737, "item_id": 46671671, "url": "https://news.ycombinator.com/item?id=46671671", "by": "basher", "date": "2026-01-18", "stance": "warn", "quote": "I tried to use the official `gemini-cli` to avoid the API costs. It is unusable for agentic work.", "excerpt": "I\u2019ve been paying *$125/mo* (promo pricing for the $250/mo tier) for Google\u2019s top-tier \"AI Ultra\" subscription for a month. I signed up because I\u2019m a power user\u2014I manage large repos, build my own MCP tools, and live in the terminal. I assumed \"Ultra\" meant \"Pro Developer.\" I was wrong. I downgraded to the $20/mo plan today, and here is why other developers should probably avoid the top tier. 1. The \"Double Pay\" Parado", "context": "power user managing large repos and building MCP tools", "use_case": "agentic coding in terminal", "switched_from": "", "story": "Google's \"AI Ultra\" Is a Storage Plan Disguised as a Dev Tool"}, {"id": 746110113, "item_id": 46631882, "url": "https://news.ycombinator.com/item?id=46631882", "by": "theshrike79", "date": "2026-01-15", "stance": "warn", "quote": "Gemini seems like an afterthought an intern did", "excerpt": "I'm pretty sure Anthropic is the only company that actually dogfoods their CLI tool. Gemini seems like an afterthought an intern did, Codex has cool features but none of them align with real-world need.", "context": "", "use_case": "CLI coding agent", "switched_from": "", "story": "Ask HN: What is best way to provide continuous context to models?"}, {"id": 746079872, "item_id": 46629992, "url": "https://news.ycombinator.com/item?id=46629992", "by": "esperent", "date": "2026-01-15", "stance": "warn", "quote": "Then Murphy's law hit me with bug after bug, all of which have been reported many times already going back over year or more.", "excerpt": "I actually defended Gemini CLI when someone said this a few days ago. Then Murphy's law hit me with bug after bug, all of which have been reported many times already going back over year or more. I keep having to totally clear out the ~/.gemini folder and then it works again for a while.", "context": "", "use_case": "CLI coding agent", "switched_from": "", "story": "Ask HN: What is best way to provide continuous context to models?"}, {"id": 744672928, "item_id": 46542058, "url": "https://news.ycombinator.com/item?id=46542058", "by": "mattarm", "date": "2026-01-08", "stance": "recommend", "quote": "", "excerpt": "Yeah, the online Gemini app is not good for long lived conversations that build up a body of decisions. The context window gets too large and things drop. What I\u2019ve learned is that once you reach that point you\u2019ve got to break that problem down into smaller pieces that the AI can work productively with. If you\u2019re about to start with Gemini-cli I recommend you look up https://github.com/github/spec-kit. It\u2019s a project", "context": "", "use_case": "AI-assisted document/code editing", "switched_from": "Gemini", "story": "Opus 4.5 is going to change everything"}, {"id": 744604512, "item_id": 46537782, "url": "https://news.ycombinator.com/item?id=46537782", "by": "mythz", "date": "2026-01-07", "stance": "warn", "quote": "Gemini CLI is too slow to be useful", "excerpt": "Gemini CLI is too slow to be useful, kind of surprised it was even offered and marketed given how painful it is to use. I thought it'd have to be damaging to the Gemini brand to get people to try it out, suffer painful UX then immediately stop using it. (Using it from Australia may also contribute to its slow perf) Antigravity was also painful to use at launch where more queries failed then succeeded, however they've", "context": "using from Australia", "use_case": "AI coding", "switched_from": "", "story": "Google Got Its Groove Back and Edged Ahead of OpenAI"}, {"id": 744514739, "item_id": 46532171, "url": "https://news.ycombinator.com/item?id=46532171", "by": "vedmakk", "date": "2026-01-07", "stance": "warn", "quote": "Gemini CLI is a pain to use - after getting used to the smoothness of Claude\nCode.", "excerpt": "I don't get the Gemini 3 hype... yes it's their first usable model, but its not even close to what Opus 4.5 and GPT 5.2 can do. Maybe on Benchmarks... but I'm forced to use Gemini at work everyday, while I use Opus 4.5 / GPT 5.2 privately every day... and Gemini is just lacking so much wit, creativity and multi-step problem solving skills compared to Opus. Not to mention that Gemini CLI is a pain to use - after getti", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Google Got Its Groove Back and Edged Ahead of OpenAI"}, {"id": 744150739, "item_id": 46509421, "url": "https://news.ycombinator.com/item?id=46509421", "by": "Abishek_Muthian", "date": "2026-01-05", "stance": "neutral", "quote": "In external agents - I have Calude Code, Codex CLI, Gemini CLI", "excerpt": "I just have a minimalist setup. Configuration - In external agents - I have Calude Code, Codex CLI, Gemini CLI and a custom agent for Qwen Coder via Llama.cpp [1] In MCP - I have fetch, brave-search, puppeteer In LLM Providers - I have configured Llama.cpp, LMStudio and OpenAI (Zed Agent can access models from any of these providers). Workflow - When I need LLM assist I mostly use just Claude Code for specific tasks,", "context": "", "use_case": "external agent in Zed", "switched_from": "", "story": "I switched from VSCode to Zed"}, {"id": 743995938, "item_id": 46499746, "url": "https://news.ycombinator.com/item?id=46499746", "by": "mark_l_watson", "date": "2026-01-05", "stance": "neutral", "quote": "preferring to separately running gemini-cli occasionally", "excerpt": "This article struck a personal chord with me: I bought a new MacBook a week ago and installed minimal software on it, specifically I did not install VSCode and I don\u2019t miss it. I use Emacs exclusively on my new laptop. I have about 40 years experience with Emacs and except for a treemacs automations, I am using my regular setup. VSCode is a great project but I just didn\u2019t feel \u201chappy\u201d while I was using it. I feel hap", "context": "40 years of Emacs experience", "use_case": "occasional LLM use", "switched_from": "", "story": "I switched from VSCode to Zed"}, {"id": 737045889, "item_id": 46065368, "url": "https://news.ycombinator.com/item?id=46065368", "by": "MarkMarine", "date": "2025-11-26", "stance": "warn", "quote": "I tried Gemini cli for about 5 seconds and was so frustrated, it\u2019s so stupid at navigation in the codebase", "excerpt": "I love the model, hate the tool. I\u2019ve taken complex stuff and given it to Gemini 3 and been impressed, but Anthropic has the killer app with Claude Code. The interplay of sonnet (a decent model) and the tools and workflow they\u2019ve got with Claude code around it supercharge the outcome. I tried Gemini cli for about 5 seconds and was so frustrated, it\u2019s so stupid at navigation in the codebase it takes 10x as long to do ", "context": "", "use_case": "coding agent", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}, {"id": 737029072, "item_id": 46064317, "url": "https://news.ycombinator.com/item?id=46064317", "by": "terrywang", "date": "2025-11-26", "stance": "recommend", "quote": "Gemini CLI is open source and being actively developed, which is cool", "excerpt": "Gemini CLI at this stage isn't good at complex coding tasks (vs. Claude Code, Codex, Cursor CLI, Qoder CLI, etc.). Mostly because of the simple ReAct loop, compounded by relatively weak tool calling capability of the Gemini 2.5 Pro model. > I haven't tried complex coding tasks using Gemini 3.0 Pro Preview yet. I reckon it won't be materially different. Gemini CLI is open source and being actively developed, which is ", "context": "", "use_case": "orchestrating tasks and workflows", "switched_from": "", "story": "Gemini CLI tips and tricks for agentic coding"}, {"id": 736169824, "item_id": 46010614, "url": "https://news.ycombinator.com/item?id=46010614", "by": "keithba", "date": "2025-11-21", "stance": "recommend", "quote": "I used Gemini CLI (once it came, AI Studio before that) with Gemini 2.5 Pro to create one with Rust and LLVM", "excerpt": "LLMs can definitely write programming languages. I used Gemini CLI (once it came, AI Studio before that) with Gemini 2.5 Pro to create one with Rust and LLVM: https://github.com/GoogleCloudPlatform/aether This was completely vibe coded - I never had to edit the code, though it was very interactive. The whole thing tool less than a month of some evenings and weekends. (Note: it\u2019s ugly on purpose, as I\u2019m playing with i", "context": "", "use_case": "vibe coding a language", "switched_from": "AI Studio", "story": "FAWK: LLMs can write a language interpreter"}, {"id": 735788498, "item_id": 45986781, "url": "https://news.ycombinator.com/item?id=45986781", "by": "freediver", "date": "2025-11-19", "stance": "warn", "quote": "Gemini CLI was never a contender (between log in/activation/rate limits - wth), will say though that Gemini CLI has the nicest terminal UI.", "excerpt": "First time that there is a worthy alternative to Claude Code. Codex Max solved a problem I had Claude Code fail multiple times. Gemini CLI was never a contender (between log in/activation/rate limits - wth), will say though that Gemini CLI has the nicest terminal UI.", "context": "", "use_case": "AI coding", "switched_from": "", "story": "Building more with GPT-5.1-Codex-Max"}]}