{"slug": "gpt-luna", "name": "GPT Luna", "category": "ai", "category_name": "AI models and assistants", "description": "OpenAI's small, low-cost GPT model tier (GPT-5.6 / GPT-6 Luna) for API and ChatGPT/Codex use.", "recommenders": 64, "commenters": 99, "warners": 7, "score": 60.535, "rising": 49.4, "first": "2026-07-12", "latest": "2026-10-29", "why": "Liked as a very cheap workhorse model for routine coding, subagents and high-volume batch tasks.", "why_is_quote": false, "has_verdict": true, "sentiment": {"recommend": 69, "neutral": 23, "compare": 8, "warn": 7}, "spark": [0, 0, 0, 0, 0, 0, 0, 0, 11, 12, 29, 17], "listed": true, "loved": ["Cost efficiency", "Routine workhorse"], "caveat": "Limited intelligence", "short_quote": "Luna is the undisputed king here at the moment and is what I use as my workhorse model.", "icon": true, "domain": "openai.com", "github_repo": null, "official_url": "https://openai.com", "pricing": "paid", "aliases": ["Luna", "GPT-5.6 Luna", "GPT 5.6 Luna", "GPT-6 Luna", "GPT 6 Luna", "GPT Luna", "OpenAI Luna", "gpt-luna", "gpt-5.6-luna", "luna"], "verdict": [{"text": "Liked as a very cheap workhorse model", "cites": [{"id": 804031328, "item_id": 50251958, "url": "https://news.ycombinator.com/item?id=50251958", "by": "rockdoe", "date": "2026-10-26", "stance": "recommend", "quote": "Luna and Sol-6.1 are crazy cost effective and all flows here have switched to them", "excerpt": "Luna and Sol-6.1 are crazy cost effective and all flows here have switched to them, so I'd say really well so far. There's very few stuff that needs Astra/Opus level intelligence. Luna 6.1 would be amazing.", "context": "", "use_case": "cost-effective workflows", "switched_from": "", "story": "GPT-6.1 Terra"}, {"id": 792652240, "item_id": 49540765, "url": "https://news.ycombinator.com/item?id=49540765", "by": "gundmc", "date": "2026-09-02", "stance": "recommend", "quote": "Luna is the undisputed king here at the moment and is what I use as my workhorse model.", "excerpt": "There are numerous benchmarks that measure cost per task, which factors out tokens entirely. Gemini 3.8 flash is significantly lower than Sol on basically all of them https://artificialanalysis.ai/#cost-tabs That said, Luna is the undisputed king here at the moment and is what I use as my workhorse model.", "context": "", "use_case": "workhorse model", "switched_from": "", "story": "Gemini 3.8 Flash and 3.8 Flash Cyber"}, {"id": 795263616, "item_id": 49703976, "url": "https://news.ycombinator.com/item?id=49703976", "by": "spockz", "date": "2026-09-14", "stance": "recommend", "quote": "To me Luna is the work horse", "excerpt": "Quite good as work horses? To me Luna is the work horse and Sol and Astra are prancing thoroughbreds. If I use Sol or Astra for anything other than curated reasoning and planning I will burn through my usage limits in an hour.", "context": "", "use_case": "general coding work", "switched_from": "", "story": "GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?"}, {"id": 792414272, "item_id": 49525892, "url": "https://news.ycombinator.com/item?id=49525892", "by": "2001zhaozhao", "date": "2026-09-01", "stance": "recommend", "quote": "Luna is insanely good for the cost and it's my go-to for high volume batch tasks now.", "excerpt": "They really should launch a new Haiku to compete with Luna imho. Luna is insanely good for the cost and it's my go-to for high volume batch tasks now.", "context": "", "use_case": "high volume batch tasks", "switched_from": "", "story": "Claude Fable 5.1 and Claude Mythos 5.1"}]}, {"text": " for routine coding, subagents and high-volume batch tasks.", "cites": [{"id": 797339746, "item_id": 49833734, "url": "https://news.ycombinator.com/item?id=49833734", "by": "anothereng", "date": "2026-09-24", "stance": "recommend", "quote": "If it is something super simple like changing a css class I use Luna.", "excerpt": "Depends on the complexity of the task. And how fast I want it solved. If I wanted solved fast I use a better model (Sol), If I can afford some time I use Terra (I use terra for most coding things). If it is something super simple like changing a css class I use Luna.", "context": "", "use_case": "simple coding tasks", "switched_from": "", "story": "GPT-6 Sol and Luna"}, {"id": 795306112, "item_id": 49706632, "url": "https://news.ycombinator.com/item?id=49706632", "by": "pkulak", "date": "2026-09-15", "stance": "recommend", "quote": "I use Luna for research/scout/test subagents", "excerpt": "Same. I use Luna for research/scout/test subagents, Sol for coordinator, Terra for delegate, and Astra high for review and simplify. Even with Astra in there, it\u2019s fresh context, and I\u2019m consistently amazed by how far I can stretch my $20 subscription with really good results.", "context": "$20 subscription user", "use_case": "research/test subagents", "switched_from": "", "story": "GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?"}, {"id": 795549795, "item_id": 49721862, "url": "https://news.ycombinator.com/item?id=49721862", "by": "pod_krad", "date": "2026-09-16", "stance": "recommend", "quote": "Terra, Luna - for implementation in simple routine cases.", "excerpt": "Astra xhigh - for specs, analysis. Sol xhigh - for orchestration and implementation in complicated cases. Terra, Luna - for implementation in simple routine cases.", "context": "", "use_case": "simple implementation", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}, {"id": 790490129, "item_id": 49405633, "url": "https://news.ycombinator.com/item?id=49405633", "by": "asats", "date": "2026-08-23", "stance": "recommend", "quote": "just paying the cheap Luna prices to openai and getting it done in a few hours", "excerpt": "Qwen3.8-27B runs at 59.5 tok/s on my M4 Max, 40-core GPU, 128 GB I use it occasionally for classification and other tasks but I wouldn't trust those smaller models with the real work and for larger data processing it's too slow, e.g. a dataset I wanted to classify would've taken 56 days on my laptop vs just paying the cheap Luna prices to openai and getting it done in a few hours.", "context": "", "use_case": "bulk dataset classification", "switched_from": "", "story": "Why your local LLM feels dumber than it is"}]}, {"text": " The most common complaint is that it reasons less well than larger models and misses obvious things,", "cites": [{"id": 800204070, "item_id": 50012754, "url": "https://news.ycombinator.com/item?id=50012754", "by": "gregwebs", "date": "2026-10-08", "stance": "warn", "quote": "Luna is cost competitive, but doesn't score as well on intelligence.", "excerpt": "I have been using DeepSeek 4.1 flash intensively for over a month. If I run it all day long it costs $1-2. Its fast. Previously I was always quickly running up to my Claude/Codex 5 hour window (on the $20/month plan). The cost savings of DeepSeek is real as shown in this article and I am using subsidized plans. DeepSeek is horrible at grilling sessions (the /grill* skills to make technical decisions). It doesn't know", "context": "", "use_case": "", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 800189264, "item_id": 50011829, "url": "https://news.ycombinator.com/item?id=50011829", "by": "pessimizer", "date": "2026-10-08", "stance": "warn", "quote": "Luna misses the obvious (and more importantly, the clearly explained) consistently.", "excerpt": "I'm no expert, but it think that it's the pricing on GPT-6 Luna. I'm also guessing that it's been underpriced just for this reason. I also don't think it's all that great, but it's definitely very cheap. If it's underpriced, it's a loss leader to sell the other models, so it actually can't be too good. I really put these things through their paces because I use them to review and work with new abstract game rules and", "context": "", "use_case": "reviewing game rules", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 792880018, "item_id": 49555001, "url": "https://news.ycombinator.com/item?id=49555001", "by": "lacoolj", "date": "2026-09-03", "stance": "warn", "quote": "I know Luna isn't reliable", "excerpt": "Jesus what are you doing that requires Sol usage so often? Terra not enough? I know Luna isn't reliable, so that's fair. Genuinely curious though, because I use Cursor daily and almost everything I do, highly complex or high volume, can be handled with Auto mode or Composer 2.5 (or Grok 4.6 High). So I have to assume you're doing something far more complex than what I am", "context": "", "use_case": "", "switched_from": "", "story": "OpenAI begins rolling out GPT-6 Astra"}, {"id": 799770321, "item_id": 49985645, "url": "https://news.ycombinator.com/item?id=49985645", "by": "Topfi", "date": "2026-10-06", "stance": "warn", "quote": "Lunas output is not really helpful here.", "excerpt": "Very task-dependent of course and mine are unique to say the least, so could see Luna being better in certain domains, even if my measurements have not shown that yet, happy for anyone to show otherwise. For what it's worth, ran every task twice on each model, most were for some UI component synthesis and charting insanity that is a bit hard to explain, but some were simple tag selection, basic noul at threshold 60%.", "context": "", "use_case": "tag selection decisions", "switched_from": "", "story": "Decisions API is in public beta"}]}, {"text": " with some saying its quality has slipped and that high-reasoning runs are slow.", "cites": [{"id": 796639232, "item_id": 49789952, "url": "https://news.ycombinator.com/item?id=49789952", "by": "jotato", "date": "2026-09-21", "stance": "warn", "quote": "over the last 2 or 3 weeks I've seen how dumb it is now", "excerpt": "Just yesterday I was thinking about gpt-5.6-luna. I made it my default model in Hermes during its fist week of launch. It was just as good as 5.5 which was my previous default. But over the last 2 or 3 weeks I've seen how dumb it is now. I have to be very explicit with it. For example, I used to be able to prompt \"Check the system logs on for....\" and it would just figure it out. Yesterday I asked \"Did $400 corporate allowance), then it's pretty obvious that Luna is small and fast, and Sol 6.1 is the big one for jobs that need significantly above average intelligence, comparatively to a human engineer. I use both constantly. Astra is really not relevant for the average dev except to do some moonshot task with a vague verification loop. Sure Sol 6.1 ", "context": "uses models daily on a modest budget", "use_case": "small fast coding tasks", "switched_from": "", "story": "GPT-6.1 Terra"}, {"id": 797681616, "item_id": 49855101, "url": "https://news.ycombinator.com/item?id=49855101", "by": "humlex", "date": "2026-09-26", "stance": "recommend", "quote": "i enjoy it a lot more if i use a super fast model with low reasoning (gpt 6 luna low effort + fast mode)", "excerpt": "I recently (2 days ago) noticed i enjoy it a lot more if i use a super fast model with low reasoning (gpt 6 luna low effort + fast mode). That way, i can still have hands on the entire time and dont have to wait around for 20minutes for it to finish up and have made a bunch of decisions without me. Also, LLMs are first and foremost excellent at reading ultra fast. Makes it excellent for summarizing and re-representin", "context": "", "use_case": "hands-on coding", "switched_from": "", "story": "How to keep enjoying programming in a world of LLMs"}, {"id": 794776288, "item_id": 49673518, "url": "https://news.ycombinator.com/item?id=49673518", "by": "oduis", "date": "2026-09-12", "stance": "recommend", "quote": "I found GPT 5.6 LUNA to be astonishingly capable.", "excerpt": "I like AI more on the short leash, giving it specific agents tasks, one at a time (centaur mode). I found GPT 5.6 LUNA to be astonishingly capable. Plus, it so fast, that it does not block my flow of throughts, like the more capable but slower models often do. And it is so cheap that I do not use Ollama local models anymore. Luna is far more capable, and so cheap that the energy prices here in Germany eat the gains o", "context": "based in Germany", "use_case": "agentic coding tasks", "switched_from": "Ollama", "story": "Ask HN: What default model do you use and why?"}]}, {"label": "Stretches plan limits", "summary": "Users say it uses little quota on Plus and Codex plans, so they rarely hit usage limits.", "quotes": [{"id": 794823600, "item_id": 49676475, "url": "https://news.ycombinator.com/item?id=49676475", "by": "nickthegreek", "date": "2026-09-12", "stance": "recommend", "quote": "luna (max reasoning). price, as it sips quota on the Plus plan and is pretty competent for my needs.", "excerpt": "luna (max reasoning). price, as it sips quota on the Plus plan and is pretty competent for my needs. else I pretty much jump to Sol (high). i find that luna can work through a 5hr quota window on a /goal without hitting 0%.", "context": "", "use_case": "default coding model", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}, {"id": 799954480, "item_id": 49997155, "url": "https://news.ycombinator.com/item?id=49997155", "by": "notatoad", "date": "2026-10-07", "stance": "recommend", "quote": "used <10% of my 5hr limit on a $100 codex plan.", "excerpt": "not haiku, but luna - last week i used it for things like \"read this historical dump of 15k support tickets and break them into categories that make sense, then propose help docs that i could write to handle the most frequent queries in each category\" used <10% of my 5hr limit on a $100 codex plan.", "context": "", "use_case": "categorizing support tickets", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 794778896, "item_id": 49673681, "url": "https://news.ycombinator.com/item?id=49673681", "by": "conradludgate", "date": "2026-09-12", "stance": "recommend", "quote": "For personal use, Luna-xhigh on the Plus subscription.", "excerpt": "For personal use, Luna-xhigh on the Plus subscription. I only ever hit my limits if I decide to use Sol or Astra for a bit more quality. For work, given we pay API pricing anyway, I've been happy experimenting with K3-medium as my default, and now I'm planning on trying GLM 5.3 as well. Sol-medium is my fallback for my work purposes but I occasionally use Opus 4.8 as an additional reviewer. I like the open weight mod", "context": "", "use_case": "personal coding", "switched_from": "", "story": "Ask HN: What default model do you use and why?"}]}, {"label": "Batch processing", "summary": "Picked for high-volume jobs such as classifying large datasets and support-ticket dumps.", "quotes": [{"id": 792414272, "item_id": 49525892, "url": "https://news.ycombinator.com/item?id=49525892", "by": "2001zhaozhao", "date": "2026-09-01", "stance": "recommend", "quote": "Luna is insanely good for the cost and it's my go-to for high volume batch tasks now.", "excerpt": "They really should launch a new Haiku to compete with Luna imho. Luna is insanely good for the cost and it's my go-to for high volume batch tasks now.", "context": "", "use_case": "high volume batch tasks", "switched_from": "", "story": "Claude Fable 5.1 and Claude Mythos 5.1"}, {"id": 790490129, "item_id": 49405633, "url": "https://news.ycombinator.com/item?id=49405633", "by": "asats", "date": "2026-08-23", "stance": "recommend", "quote": "just paying the cheap Luna prices to openai and getting it done in a few hours", "excerpt": "Qwen3.8-27B runs at 59.5 tok/s on my M4 Max, 40-core GPU, 128 GB I use it occasionally for classification and other tasks but I wouldn't trust those smaller models with the real work and for larger data processing it's too slow, e.g. a dataset I wanted to classify would've taken 56 days on my laptop vs just paying the cheap Luna prices to openai and getting it done in a few hours.", "context": "", "use_case": "bulk dataset classification", "switched_from": "", "story": "Why your local LLM feels dumber than it is"}, {"id": 799954480, "item_id": 49997155, "url": "https://news.ycombinator.com/item?id=49997155", "by": "notatoad", "date": "2026-10-07", "stance": "recommend", "quote": "used <10% of my 5hr limit on a $100 codex plan.", "excerpt": "not haiku, but luna - last week i used it for things like \"read this historical dump of 15k support tickets and break them into categories that make sense, then propose help docs that i could write to handle the most frequent queries in each category\" used <10% of my 5hr limit on a $100 codex plan.", "context": "", "use_case": "categorizing support tickets", "switched_from": "", "story": "Claude Haiku 5.5"}]}, {"label": "Code review", "summary": "Some find it good at focused code review when the task is narrowly scoped and split across reviewers.", "quotes": [{"id": 795271328, "item_id": 49704458, "url": "https://news.ycombinator.com/item?id=49704458", "by": "CharlieDigital", "date": "2026-09-14", "stance": "recommend", "quote": "I found Luna and even 5.4-mini to be quite good at code review", "excerpt": "I found Luna and even 5.4-mini to be quite good at code review provided a few things: 1. Run it in multiple cycles, only on the diff, and only emit a few findings at a time. 2. Give it a memory so each cycle, it knows the previous finding to check if it's been fixed. 3. Give it access to canonical docs that encode your human reviewer heuristics. I exposed these as tool calls so they could be tracked via telemetry. 4.", "context": "", "use_case": "code review", "switched_from": "", "story": "GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?"}, {"id": 795289569, "item_id": 49705598, "url": "https://news.ycombinator.com/item?id=49705598", "by": "jbellis", "date": "2026-09-14", "stance": "recommend", "quote": "six subagents (I recommend Luna)", "excerpt": "For the mj adversarial review we use a configurable coordinator (I recommend Fable) with six subagents (I recommend Luna) for specific specialties like code complexity. Narrowing down what a smaller model like Luna needs to look for / care about helps them do better work; then the larger model synthesizes and fills in any gaps. https://blog.brokk.ai/mjolnir-automated-cross-vendor-adversarial-review/", "context": "", "use_case": "code review subagents", "switched_from": "", "story": "GPT-5.6 Luna vs. GPT-6 Astra: Is a $1.20 Model Good Enough for Code Review?"}]}], "caveats": [{"label": "Limited intelligence", "summary": "Several commenters say it scores lower on intelligence, misses obvious points and is unreliable for harder work.", "quotes": [{"id": 800204070, "item_id": 50012754, "url": "https://news.ycombinator.com/item?id=50012754", "by": "gregwebs", "date": "2026-10-08", "stance": "warn", "quote": "Luna is cost competitive, but doesn't score as well on intelligence.", "excerpt": "I have been using DeepSeek 4.1 flash intensively for over a month. If I run it all day long it costs $1-2. Its fast. Previously I was always quickly running up to my Claude/Codex 5 hour window (on the $20/month plan). The cost savings of DeepSeek is real as shown in this article and I am using subsidized plans. DeepSeek is horrible at grilling sessions (the /grill* skills to make technical decisions). It doesn't know", "context": "", "use_case": "", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 800189264, "item_id": 50011829, "url": "https://news.ycombinator.com/item?id=50011829", "by": "pessimizer", "date": "2026-10-08", "stance": "warn", "quote": "Luna misses the obvious (and more importantly, the clearly explained) consistently.", "excerpt": "I'm no expert, but it think that it's the pricing on GPT-6 Luna. I'm also guessing that it's been underpriced just for this reason. I also don't think it's all that great, but it's definitely very cheap. If it's underpriced, it's a loss leader to sell the other models, so it actually can't be too good. I really put these things through their paces because I use them to review and work with new abstract game rules and", "context": "", "use_case": "reviewing game rules", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 792880018, "item_id": 49555001, "url": "https://news.ycombinator.com/item?id=49555001", "by": "lacoolj", "date": "2026-09-03", "stance": "warn", "quote": "I know Luna isn't reliable", "excerpt": "Jesus what are you doing that requires Sol usage so often? Terra not enough? I know Luna isn't reliable, so that's fair. Genuinely curious though, because I use Cursor daily and almost everything I do, highly complex or high volume, can be handled with Auto mode or Composer 2.5 (or Grok 4.6 High). So I have to assume you're doing something far more complex than what I am", "context": "", "use_case": "", "switched_from": "", "story": "OpenAI begins rolling out GPT-6 Astra"}]}, {"label": "Quality regression", "summary": "Some report the model seemed to get dumber over time and users are upset.", "quotes": [{"id": 796639232, "item_id": 49789952, "url": "https://news.ycombinator.com/item?id=49789952", "by": "jotato", "date": "2026-09-21", "stance": "warn", "quote": "over the last 2 or 3 weeks I've seen how dumb it is now", "excerpt": "Just yesterday I was thinking about gpt-5.6-luna. I made it my default model in Hermes during its fist week of launch. It was just as good as 5.5 which was my previous default. But over the last 2 or 3 weeks I've seen how dumb it is now. I have to be very explicit with it. For example, I used to be able to prompt \"Check the system logs on for....\" and it would just figure it out. Yesterday I asked \"Did for....\" and it would just figure it out. Yesterday I asked \"Did $400 corporate allowance), then it's pretty obvious that Luna is small and fast, and Sol 6.1 is the big one for jobs that need significantly above average intelligence, comparatively to a human engineer. I use both constantly. Astra is really not relevant for the average dev except to do some moonshot task with a vague verification loop. Sure Sol 6.1 ", "context": "uses models daily on a modest budget", "use_case": "small fast coding tasks", "switched_from": "", "story": "GPT-6.1 Terra"}, {"id": 803980112, "item_id": 50248757, "url": "https://news.ycombinator.com/item?id=50248757", "by": "eknkc", "date": "2026-10-26", "stance": "recommend", "quote": "On the API we use a lot of Luna (cost efficient)", "excerpt": "On the API we use a lot of Luna (cost efficient) and sometimes Terra level models. I think we never use a larger model. These are models that our app runtimes depend on. On subscription for coding and stuff I am basically on Astra Ultra all day long.", "context": "", "use_case": "app runtime API calls", "switched_from": "", "story": "GPT-6.1 Terra"}, {"id": 803494438, "item_id": 50218402, "url": "https://news.ycombinator.com/item?id=50218402", "by": "sail0rm00n", "date": "2026-10-24", "stance": "recommend", "quote": "Gemini Flash 3.8, DeepSeek Flash, and Luna are still up there as favorites", "excerpt": "As tribute, I bit the bullet on the $40 plan to test out K3.1 Pro. Still burning through the 5-hour rate limit way too quickly, though. Output quality seems solid enough, but the value will probably make a lot more sense via API pricing for smaller projects. My current tier list based on what I\u2019ve been testing: 1. Opus 5.5 2. Astra 3. Argon 4. Kimi K3.1 Pro (For lightweight/daily driver tasks, Gemini Flash 3.8, DeepS", "context": "", "use_case": "lightweight daily tasks", "switched_from": "", "story": "Kimi K3.1 Pro and K3.1 Pro Highspeed"}, {"id": 803262144, "item_id": 50203884, "url": "https://news.ycombinator.com/item?id=50203884", "by": "dawnerd", "date": "2026-10-23", "stance": "compare", "quote": "Luna seems more willing than Sol", "excerpt": "Not very interesting, just being vague to not give it any context that it might be doing security research. Treating it like a game also works fairly well. It's hard to get more detailed about how since I use it for work and don't want it patched out, ya know? Splitting it up between turns, like Luna seems more willing than Sol. If you get blocked better to take what context you have and start a new chat and try to r", "context": "Uses it for work", "use_case": "security research prompting", "switched_from": "", "story": "Jailbreaking the Switch 2"}, {"id": 802763589, "item_id": 50172724, "url": "https://news.ycombinator.com/item?id=50172724", "by": "CharlieDigital", "date": "2026-10-20", "stance": "recommend", "quote": "Coding is handled by Haiku 5.5/GPT 6 Luna/DeepSeek v4.1 Flash.", "excerpt": "I use the frontier models only in plan mode; coding is held out to cheaper models. I run Paseo as an orchestrator with agent profiles. \"Distinguished Member of Technical Staff\" -> Fable, plan mode. \"End-User Customer Advocate\" -> Astra, plan mode. A few others. Basically, these agents (with profile prompts) can only be \"consulted\" for input on design elements; it's OK if they produce a bit of verbosity here. This has", "context": "", "use_case": "coding", "switched_from": "", "story": "SlopCodeBench 10/2026 \u2013 Opus 5.5, Astra wins overall, xhigh considered harmful"}, {"id": 802555378, "item_id": 50159711, "url": "https://news.ycombinator.com/item?id=50159711", "by": "simonw", "date": "2026-10-20", "stance": "recommend", "quote": "Claude Haiku 5.5 and GPT-6 Luna - both excellent inexpensive models", "excerpt": "There are a few problems with using OpenRouter's public statistics to make statements about the wider field: - It only tracks companies that have decided to use OpenRouter to help them switch models - which biases towards companies that like open weights. Anyone using OpenAI or Anthropic directly is much less likely to be represented. - OpenRouter don't share the number of unique users represented in those numbers. I", "context": "", "use_case": "", "switched_from": "", "story": "Token-maxxing is dead, OpenRouter's token usage plateaus"}, {"id": 802082049, "item_id": 50130128, "url": "https://news.ycombinator.com/item?id=50130128", "by": "clx75", "date": "2026-10-17", "stance": "recommend", "quote": "My daily driver was Luna, with occasional elevation to Terra or Sol", "excerpt": "Probably a week ago someone posted a thread with the title \"Why isn't the industry freaking out about DeepSeek 4.1 Flash?\" I have been intrigued about Chinese models for a long time, but never tried anything except an early Qwen. My daily driver was Luna, with occasional elevation to Terra or Sol - all on 5.6 as I found the output of 6 jarringly incomprehensible. I still had 95 USD on my OpenRouter account, so I thou", "context": "", "use_case": "planning agent", "switched_from": "", "story": "DeepSeek-v4.1-Flash-1017"}, {"id": 800968977, "item_id": 50060561, "url": "https://news.ycombinator.com/item?id=50060561", "by": "vishnukool", "date": "2026-10-12", "stance": "neutral", "quote": "We use OpenAI Luna for the agent37/default model", "excerpt": "We use OpenAI Luna for the agent37/default model for that. Yeah in addition to sandbox we also have managed MCPs and our own LLM gateway to make building hosting agent harnesses easier. Most of our customers host hermes, opencode, claude code etc. and needed an easier setup so we ended up offering our own LLM gateway for it, ready to use with the sandbox", "context": "", "use_case": "free tier default model", "switched_from": "", "story": "Launch HN: Agent37 (YC F26) \u2013 Low-cost agent sandboxes for running agents"}, {"id": 800914304, "item_id": 50057144, "url": "https://news.ycombinator.com/item?id=50057144", "by": "igleria", "date": "2026-10-12", "stance": "recommend", "quote": "I love luna but unfortunately it needs to be much, much faster in tokens per second", "excerpt": "I love luna but unfortunately it needs to be much, much faster in tokens per second to compensate for lack of intelligence.", "context": "", "use_case": "programming", "switched_from": "", "story": "The gap in shared understanding of LLM capability is widening"}, {"id": 800862960, "item_id": 50053935, "url": "https://news.ycombinator.com/item?id=50053935", "by": "rwyinuse", "date": "2026-10-12", "stance": "recommend", "quote": "Even cheap models like GPT 6 Luna do decent job in programming these days.", "excerpt": "Even cheap models like GPT 6 Luna do decent job in programming these days. Token prices would need to rise A LOT to make companies return to writing code by hand, instead of these models. My company has largely switched to Luna after Microsoft raised prices for Copilot.", "context": "company switched to it", "use_case": "programming", "switched_from": "Copilot", "story": "The gap in shared understanding of LLM capability is widening"}, {"id": 800429938, "item_id": 50026871, "url": "https://news.ycombinator.com/item?id=50026871", "by": "361994752", "date": "2026-10-09", "stance": "neutral", "quote": "delegate large portion of them to gpt-luna", "excerpt": "I double checked the publish date before writing down this comment. Because I use the same harness (Opencode, to be specific) as the author, and some of the features are just right there. Like opencode can start multiple subagents for different tasks in parallel. Also I usually ask the main model (e.g. opus 5.5) to pick subagent models for me, and it has no problem identify the difficulty of work and delegate large p", "context": "", "use_case": "subagent model", "switched_from": "", "story": "Why are coding agents so dumb?"}, {"id": 800403123, "item_id": 50025195, "url": "https://news.ycombinator.com/item?id=50025195", "by": "aftbit", "date": "2026-10-09", "stance": "neutral", "quote": "I often push down to Terra or Luna", "excerpt": "I have had good luck rotating between the three tiers of GPT 5.6 with occasional jumps up to Astra or Fable. Most of my work is with GPT-5.6-Sol. Simple tasks like data extraction or trivial refactors (rename this variable etc) I often push down to Terra or Luna, or even to self-hosted Gemma4:31B. Very tricky stuff, like planning a new feature, design review, or code review of a complex change across multiple reposit", "context": "", "use_case": "simple tasks", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 800234273, "item_id": 50014642, "url": "https://news.ycombinator.com/item?id=50014642", "by": "ltbarcly3", "date": "2026-10-09", "stance": "recommend", "quote": "my personal experience is Luna is better for what I do, and Luna is cheaper.", "excerpt": "DeepSeek 4.1 Flash kindof sucks. I used it a bunch and it kindof sucks. I don't know if they are gaming benchmarks or what. Luna is on par in benchmarks and my personal experience is Luna is better for what I do, and Luna is cheaper. Comparing Deepseek 4.1 flash to Opus is just ludicrous. https://artificialanalysis.ai/models/releases/comparisons?compare=gpt-6-luna%2Cdeepseek-v4-1-flash", "context": "", "use_case": "", "switched_from": "DeepSeek 4.1 Flash", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 800204070, "item_id": 50012754, "url": "https://news.ycombinator.com/item?id=50012754", "by": "gregwebs", "date": "2026-10-08", "stance": "warn", "quote": "Luna is cost competitive, but doesn't score as well on intelligence.", "excerpt": "I have been using DeepSeek 4.1 flash intensively for over a month. If I run it all day long it costs $1-2. Its fast. Previously I was always quickly running up to my Claude/Codex 5 hour window (on the $20/month plan). The cost savings of DeepSeek is real as shown in this article and I am using subsidized plans. DeepSeek is horrible at grilling sessions (the /grill* skills to make technical decisions). It doesn't know", "context": "", "use_case": "", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 800189264, "item_id": 50011829, "url": "https://news.ycombinator.com/item?id=50011829", "by": "pessimizer", "date": "2026-10-08", "stance": "warn", "quote": "Luna misses the obvious (and more importantly, the clearly explained) consistently.", "excerpt": "I'm no expert, but it think that it's the pricing on GPT-6 Luna. I'm also guessing that it's been underpriced just for this reason. I also don't think it's all that great, but it's definitely very cheap. If it's underpriced, it's a loss leader to sell the other models, so it actually can't be too good. I really put these things through their paces because I use them to review and work with new abstract game rules and", "context": "", "use_case": "reviewing game rules", "switched_from": "", "story": "Why isn't the industry freaking out about DeepSeek 4.1 Flash?"}, {"id": 800071280, "item_id": 50004455, "url": "https://news.ycombinator.com/item?id=50004455", "by": "zapnuk", "date": "2026-10-08", "stance": "recommend", "quote": "luna is so unbelievable cheap.", "excerpt": "At scale this make a difference. For me/us not that much. At my work we currently don't have a subscription and use API based billing. Some time ago the price for my usage was about 30-50\u20ac per day when we used opus for about everything. With luna+sol its about 10\u20ac for sol to plan or debug and 5\u20ac for luna to implement. At that point it doesn't quite matter to me if the cheap anthropic model is 20-50% more expensive co", "context": "uses API billing at work", "use_case": "implementation from a rough plan", "switched_from": "Opus", "story": "Claude Haiku 5.5"}, {"id": 800071155, "item_id": 50004447, "url": "https://news.ycombinator.com/item?id=50004447", "by": "jeremyjh", "date": "2026-10-08", "stance": "neutral", "quote": "Luna-max for planned coding and general tasks.", "excerpt": "I use oh-my-pi (omp.sh) - mostly with stock settings and skills but its a \"fully loaded\" harness that you don't really have to add anything to. I do change a couple of things: I set it to prefer subagents, and I enable rewind. It is crazy that rewind is not a default, it saves a TON of context - if the agent goes down a crazy path that burns a lot of tokens it can rewind to an earlier checkpoint with the exploration ", "context": "", "use_case": "planned coding", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 799991169, "item_id": 49999448, "url": "https://news.ycombinator.com/item?id=49999448", "by": "jeremyjh", "date": "2026-10-07", "stance": "recommend", "quote": "I use Luna for this day in and out and its excellent", "excerpt": "If you can't get any coding done with 100K context that is either a broken model, a broken harness or a skill issue. I would mostly use Haiku in task or explorer subagents. I'm not saying I stay under that on every task, but I do have quite a few sessions that cap out well below that, so that price difference would be very meaningful. I use Luna for this day in and out and its excellent - if Haiku is that much better", "context": "", "use_case": "subagent tasks", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 799954480, "item_id": 49997155, "url": "https://news.ycombinator.com/item?id=49997155", "by": "notatoad", "date": "2026-10-07", "stance": "recommend", "quote": "used <10% of my 5hr limit on a $100 codex plan.", "excerpt": "not haiku, but luna - last week i used it for things like \"read this historical dump of 15k support tickets and break them into categories that make sense, then propose help docs that i could write to handle the most frequent queries in each category\" used <10% of my 5hr limit on a $100 codex plan.", "context": "", "use_case": "categorizing support tickets", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 799953683, "item_id": 49997105, "url": "https://news.ycombinator.com/item?id=49997105", "by": "wyrdcurt", "date": "2026-10-07", "stance": "recommend", "quote": "I use Luna for certain tasks where speed is more valuable than performance", "excerpt": "Some people/organizations are ideologically opposed to using Chinese models. Not me, I use GLM-5.3-Flash for almost everything (the subscription-subsidized pricing on a legacy Z.ai plan makes it the best value model by a wide margin), along with some MiMo and DeepSeek. Still, I use Luna for certain tasks where speed is more valuable than performance; I can see this new Haiku displacing Luna for those. If you mean Hai", "context": "", "use_case": "speed-sensitive tasks", "switched_from": "", "story": "Claude Haiku 5.5"}, {"id": 799770321, "item_id": 49985645, "url": "https://news.ycombinator.com/item?id=49985645", "by": "Topfi", "date": "2026-10-06", "stance": "warn", "quote": "Lunas output is not really helpful here.", "excerpt": "Very task-dependent of course and mine are unique to say the least, so could see Luna being better in certain domains, even if my measurements have not shown that yet, happy for anyone to show otherwise. For what it's worth, ran every task twice on each model, most were for some UI component synthesis and charting insanity that is a bit hard to explain, but some were simple tag selection, basic noul at threshold 60%.", "context": "", "use_case": "tag selection decisions", "switched_from": "", "story": "Decisions API is in public beta"}, {"id": 799767393, "item_id": 49985462, "url": "https://news.ycombinator.com/item?id=49985462", "by": "nostrebored", "date": "2026-10-06", "stance": "compare", "quote": "i would be shocked if luna decides is less generally capable than jev", "excerpt": "i would be shocked if luna decides is less generally capable than jev. jev has failed to understand any novel domain I've given it. i have found that i use it only when \"some data is better than no data\"", "context": "", "use_case": "decisions API", "switched_from": "", "story": "Decisions API is in public beta"}, {"id": 799760562, "item_id": 49985035, "url": "https://news.ycombinator.com/item?id=49985035", "by": "simonw", "date": "2026-10-06", "stance": "neutral", "quote": "", "excerpt": "curl https://api.openai.com/v1/decisions \\ -H \"Authorization: Bearer $(llm keys get openai)\" \\ -H \"Content-Type: application/json\" \\ --data ' { \"model\": \"gpt-6-luna\", \"input\": [{ \"role\": \"user\", \"content\": [ {\"type\": \"input_text\", \"text\": \"I am angry about the new product feature\"} ] }], \"questions\": [{ \"type\": \"predicate\", \"name\": \"complaint\", \"instructions\": \"Is this a complaint?\" }, { \"type\": \"predicate\", \"name\": ", "context": "", "use_case": "classifying text", "switched_from": "", "story": "Decisions API is in public beta"}, {"id": 799631778, "item_id": 49976986, "url": "https://news.ycombinator.com/item?id=49976986", "by": "nattadev", "date": "2026-10-06", "stance": "recommend", "quote": "Luna as it was cheap and more than powerful enough", "excerpt": "Hi all! I'm new to the community but wanted to show off a project I'm proud of. This is called Natta Toolkit, and it was meant to be a control center/routing method for tools I use at work and personally. I made it open source, because I really wanted anyone to be able to stress test it and give feedback. The goal I was going for was to have a single route to local scripts that otherwise needed aliases, direct calls ", "context": "", "use_case": "natural language routing", "switched_from": "OpenJev", "story": "Show HN: A Deterministic-First Local Control Center Leveraging AI for Routing"}, {"id": 798907425, "item_id": 49931714, "url": "https://news.ycombinator.com/item?id=49931714", "by": "skapadia", "date": "2026-10-02", "stance": "recommend", "quote": "cheaper models like the GPT Luna family", "excerpt": "I use pi + cheaper models like the GPT Luna family, for loop engineering", "context": "", "use_case": "loop engineering", "switched_from": "", "story": "Pi 1.0"}, {"id": 798384531, "item_id": 49899033, "url": "https://news.ycombinator.com/item?id=49899033", "by": "jjcm", "date": "2026-09-29", "stance": "recommend", "quote": "I think Luna is a fantastic driver for these", "excerpt": "There's a lot of negativity in here for Dots. I've been a pretty heavy user of Grok Bot, and here are a few thoughts a long the positive line. 1. Collaboration between always-on agents is a really, really powerful thing. It allows for domain-specific expertise that doesn't overload the context window, while still allowing for access to knowledge if they need it. 2. Domain-specific always on agents creates a good barr", "context": "", "use_case": "model driving always-on agents", "switched_from": "", "story": "Dots: Always-on agents"}, {"id": 798302641, "item_id": 49893915, "url": "https://news.ycombinator.com/item?id=49893915", "by": "dools", "date": "2026-09-29", "stance": "compare", "quote": "GPT 6 Luna closed the gap significantly for sure", "excerpt": "I go for weeks without topping up my Deepseek balance, and have to top up my OpenAI balance at least every week. GPT 6 Luna closed the gap significantly for sure (it seems to be about twice as expensive as DS v4.1 Flash), but Deepseek v4.1 Flash is still the best value model and capable enough for almost everything I need to do. Sometimes if it's babbling or can't nail down a solution I switch to Sol for one prompt, ", "context": "", "use_case": "", "switched_from": "", "story": "Sonnet 5.5"}, {"id": 798177714, "item_id": 49886107, "url": "https://news.ycombinator.com/item?id=49886107", "by": "cromka", "date": "2026-09-28", "stance": "warn", "quote": "Don't get me started on Luna.", "excerpt": "Not even GPT6 Sol cannot match DeepSeek 4.1 in my work, with outrageous bugs. Don't get me started on Luna. From my today's session with Sol: - it actually failed to correctly understand a simple English grammar and logical implication of it, then when challenged it admitted its mistake but couldn't explain why it made it. - for the code I am working on, I asked to create two PRs for the two small features (couple li", "context": "", "use_case": "", "switched_from": "", "story": "Sonnet 5.5"}, {"id": 798092533, "item_id": 49880783, "url": "https://news.ycombinator.com/item?id=49880783", "by": "caidehen", "date": "2026-09-28", "stance": "neutral", "quote": "luna implemented the project", "excerpt": "Agree with the author. I used to write code by hand, literally hand, I used vanilla vim with only syntax highlight and line number enabled, no more. So I know I can write code. I wrote C and Python most. I used to vibe code, I am still vibe coding, multiple projects at same time. I use opus, astra, luna, deepseek v4.1 flash, I use claude code, codex, pi. I vibed a project to manage my coding agents' sub agents, skill", "context": "", "use_case": "implementation", "switched_from": "", "story": "Coding is not solved"}, {"id": 797994050, "item_id": 49874628, "url": "https://news.ycombinator.com/item?id=49874628", "by": "Gareth321", "date": "2026-09-28", "stance": "warn", "quote": "I've never seen users as upset about GPT-6 Sol/Luna than I have right now.", "excerpt": "They're impossible to objectively track by design and intent. Intelligence is such a nebulous target that one can find any number of metrics to support any premise: that the model is smarter, and that the model is dumber. We use benchmarks to attempt to standardise comparisons but these are quickly ingested into the training data and then become effectively useless. You might have heard the term \"bench-maxed.\" Meanin", "context": "", "use_case": "", "switched_from": "", "story": "Ember-1"}, {"id": 797896770, "item_id": 49868548, "url": "https://news.ycombinator.com/item?id=49868548", "by": "nikhilkxmar", "date": "2026-09-27", "stance": "neutral", "quote": "for the rest of the stuff I use sonnet 5 / gpt luna", "excerpt": "I usually use Opus 5.5 for coding stuff and for the rest of the stuff I use sonnet 5 / gpt luna doesn't really matter", "context": "", "use_case": "general tasks", "switched_from": "", "story": "Ask HN: Which model do you use for work?"}, {"id": 797761057, "item_id": 49860066, "url": "https://news.ycombinator.com/item?id=49860066", "by": "Inozem", "date": "2026-09-26", "stance": "neutral", "quote": "I use gpt-6-sol for planing and luna for writing code", "excerpt": "I use gpt-6-sol for planing and luna for writing code", "context": "", "use_case": "writing code", "switched_from": "", "story": "Ask HN: Which model do you use for work?"}, {"id": 797681616, "item_id": 49855101, "url": "https://news.ycombinator.com/item?id=49855101", "by": "humlex", "date": "2026-09-26", "stance": "recommend", "quote": "i enjoy it a lot more if i use a super fast model with low reasoning (gpt 6 luna low effort + fast mode)", "excerpt": "I recently (2 days ago) noticed i enjoy it a lot more if i use a super fast model with low reasoning (gpt 6 luna low effort + fast mode). That way, i can still have hands on the entire time and dont have to wait around for 20minutes for it to finish up and have made a bunch of decisions without me. Also, LLMs are first and foremost excellent at reading ultra fast. Makes it excellent for summarizing and re-representin", "context": "", "use_case": "hands-on coding", "switched_from": "", "story": "How to keep enjoying programming in a world of LLMs"}, {"id": 797581120, "item_id": 49848820, "url": "https://news.ycombinator.com/item?id=49848820", "by": "blinded", "date": "2026-09-25", "stance": "recommend", "quote": "luna for day to day tasks", "excerpt": "luna for day to day tasks. opus for design / specs.", "context": "", "use_case": "day to day tasks", "switched_from": "", "story": "Ask HN: Which model do you use for work?"}, {"id": 797575472, "item_id": 49848467, "url": "https://news.ycombinator.com/item?id=49848467", "by": "cs1996", "date": "2026-09-25", "stance": "recommend", "quote": "in codex - luna high for 80% of everything.", "excerpt": "in codex - luna high for 80% of everything. This keeps me within the $100 a month plan. Then every so often Sol 6 high or even Astro xhigh but rarely. Also a lot of chat on xhigh or astro because that's billed seperate so I can have chat do a code review of what luna high just did and paste it to luna. Luna is always so embarrassed and agrees with the review.", "context": "", "use_case": "daily coding", "switched_from": "", "story": "Ask HN: Which model do you use for work?"}, {"id": 797400371, "item_id": 49837523, "url": "https://news.ycombinator.com/item?id=49837523", "by": "sznio", "date": "2026-09-24", "stance": "compare", "quote": "Luna doesn't feel more capable nor faster than Qwen3.8", "excerpt": "Nothing, because current Haiku is useless. But I use Qwen3.6 35B-A3B and it's both much faster and much more capable than Haiku. And Qwen3.8 27B can do so much more, decently replaces Sonnet, albeit not as fast. Luna doesn't feel more capable nor faster than Qwen3.8. I think Anthropic could surprise us with some more innovation in this niche.", "context": "", "use_case": "", "switched_from": "", "story": "Claude Opus 5.5"}, {"id": 797339746, "item_id": 49833734, "url": "https://news.ycombinator.com/item?id=49833734", "by": "anothereng", "date": "2026-09-24", "stance": "recommend", "quote": "If it is something super simple like changing a css class I use Luna.", "excerpt": "Depends on the complexity of the task. And how fast I want it solved. If I wanted solved fast I use a better model (Sol), If I can afford some time I use Terra (I use terra for most coding things). If it is something super simple like changing a css class I use Luna.", "context": "", "use_case": "simple coding tasks", "switched_from": "", "story": "GPT-6 Sol and Luna"}, {"id": 796952720, "item_id": 49809545, "url": "https://news.ycombinator.com/item?id=49809545", "by": "azuanrb", "date": "2026-09-22", "stance": "recommend", "quote": "Luna xhigh is more than good enough for me", "excerpt": "Terra is in a weird spot for me. I used to run it as my main driver at medium/high, but after Luna's price drop and some experimenting, I switched to Luna xhigh. If I need extra juice, I just use Sol. Intelligence-wise, Luna xhigh is more than good enough for me. Speed is the only downside. Terra/Sol might be similarly intelligent, but they can get things done faster.", "context": "", "use_case": "main coding model", "switched_from": "Terra", "story": "GPT-6 Sol and Luna"}, {"id": 796639232, "item_id": 49789952, "url": "https://news.ycombinator.com/item?id=49789952", "by": "jotato", "date": "2026-09-21", "stance": "warn", "quote": "over the last 2 or 3 weeks I've seen how dumb it is now", "excerpt": "Just yesterday I was thinking about gpt-5.6-luna. I made it my default model in Hermes during its fist week of launch. It was just as good as 5.5 which was my previous default. But over the last 2 or 3 weeks I've seen how dumb it is now. I have to be very explicit with it. For example, I used to be able to prompt \"Check the system logs on for....\" and it would just figure it out. Yesterday I asked \"Did