Ollama
Tool for downloading and running open-weight large language models locally, with an optional cloud plan.
Liked as an easy way to run open models like Qwen and Gemma locally and privately,12345 and for a $20/month Ollama Cloud plan that users find generous.6789 The most common complaint is that it takes llama.cpp's work without proper attribution,10111213 and many say it lags llama.cpp on speed and features.14151617
Why people recommend it
Cloud plan value · 3
Several users say the $20/month Ollama Cloud plan gives access to large open models and they rarely hit usage limits.
“The $20/mo tier is very generous, in my experience.”
“I never hit quota limits.”
Show 1 more
“Never reached the 5 hours usage limits of the $20 plan (on side projects)”
Local models · 3
Users report running Qwen, Gemma and GPT-OSS models well on their own GPUs and Macs.
“Runs pretty well with Ollama on the Qwen models.”
“ollama to run qwen 3.8 27b on my single 7900 XTX”
Show 1 more
“For my local work I use Ollama. (M4 Max 128GB)”
Easy setup · 3
Often described as an easy on-ramp that just works, much simpler to use than llama.cpp.
“Download ollama and give it a try. It doesn't have to be that one but it's an easy on-ramp.”
Show 1 more
“I use Ollama (CLI and their npm as a dev) + HF because it just works.”
Privacy · 3
Used to keep prompts and data on your own machine, with nothing sent to third parties.
“It runs entirely on your machine via Ollama (I recommend gemma3:27b)”
“using Ollama (runs 100% locally)”
Show 1 more
“It runs Ollama + Open WebUI with GPT-OSS 120b as well as GLM4.5-Air”
Developer API · 3
Developers build tools on top of its API and npm package.
“For consistent LLM behaviour, you can use ollama api with your model of choice to generate.”
“I use Ollama (CLI and their npm as a dev) + HF because it just works.”
Show 1 more
“Summaries via OpenAI or local models (I use Gemma 4B on Ollama)”
Caveats and dissent
Attribution concerns · 3
Critics say it rebadges llama.cpp work without giving proper credit and recommend alternatives.
“Highly recommend just dropping Ollama.”
“Weren't people pissed off at them for a variety of reasons? Not giving attribution etc.”
Show 1 more
“Ollama is a vampire on the culture and their demise cannot come soon enough.”
Lags llama.cpp · 3
Users report better performance and missing features, such as MoE CPU offloading, when using llama.cpp directly.
“on this VC funded joke it would be simply unusable”
“found a lot of performance improvement over ollama”
Show 1 more
“Ollama sucks, they dropped ROCm support for my card out of the blue”
Bugs · 2
Users report bugs with reasoning and streaming, and a prompt injection issue that only Ollama's developers can fix.
“due to some active bugs in ollama related to reasoning and streaming”
“for tools like Ollama there is no solution except of to fix it by Ollama developers”
Small default context · 1
It defaults to a very small context window, which can break agent harnesses.
Model storage lock-in · 1
Models are stored as hashed blobs, so other tools cannot easily reuse the same GGUF files.
“The fact that I can’t easily point another tool at the same GGUF file is a dealbreaker for me.”
Monetization worries · 3
Some users left after the team started monetizing, and others call it a VC-funded push toward its cloud platform.
“I got spooked when the Ollama team started monetizing”
“Ollama sucks, they dropped ROCm support for my card out of the blue”
Show 1 more
“on this VC funded joke it would be simply unusable”
Often compared with
14 moved from Ollama to llama.cpp
1 moved from Ollama to LM Studio
Who recommends it in their own words
From the makers not counted in the ranking
“Ollama for inference.”
Threads where it came up
- Ask HN: Who uses open LLMs and coding assistants locally? Share setup and laptop 7 mentions
- Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding? 5 mentions
- Stop Using Ollama 4 mentions
- Ask HN: What are you working on? (September 2026) 3 mentions
- Which tools do Claude, Codex and Cursor choose? We measured 17k runs to find out 3 mentions
- Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model 3 mentions
Also written as Ollama Cloud, ollama. Data: ollama.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 54
“I see the arrow on the Ollama icon telling me I have to relaunch it to apply updates.”
“Been using GLM 5.3 Flash via Ollama Cloud basic subscription”
“I used OpenCode to help me install ollama, Qwen local LLM, and connect OpenCode to it.”
“I experiment a lot with models I can run locally via ollama”
“I used GLM-5.3 as usual from Ollama cloud”
“currently building with qwen3.5 via Ollama”
“ollama defaults to incredibly tiny amounts of context”
“Here I am using it as frontend to Jupyter and ollama [2].”
“you shall use CC, serve it via ollama”
“I pay for the $20/month Ollama plan”
“I'm always excited to run models on my local machine. I use Ollama for that.”
“Zerfoo ran Gemma 3 1B at 235 tok/s in one recorded benchmark versus 188 tok/s with Ollama 0.17.7”
“When I use Claude Code locally with Ollama/Quen”
“Download Ollama (you can install the models themselves onto a USB drive so you don't fill up your hard drive with LLMs)”
“ollama to run qwen 3.8 27b on my single 7900 XTX”
“using a local `ollama` model”
“why should I not just use those directly if I am focused on local first models”
“why not we train a local ollama model 1.5 b parameters”
“It can directly connect to your local Ollama instance for inference.”
“it is so cheap that I do not use Ollama local models anymore”
“I build AI agents that run on the customer's own machine under Ollama”
“LLM chat: Ollama, Open WebUI”
“More ethical alternatives have caught up including llama.cpp”
“I recommend llama.cpp over ollama.”
“Weren't people pissed off at them for a variety of reasons? Not giving attribution etc.”
“Download ollama and give it a try. It doesn't have to be that one but it's an easy on-ramp.”
“built a personal meta-harness on top of agy, cc, codex, kiro-cli, ollama”
“Prior to two weeks ago, I was just using Pi and Ollama.”
“AI analysis happen locally (using ollama integration)”
“The AI side currently supports local models through Ollama”
“How is Ollama pivoting to a "less open" direction? Everything is MIT licensed.”
“Ollama sucks, they dropped ROCm support for my card out of the blue”
“Ollama released an update allowing it to integrate with Claude”
“Runs on top of local Ollama”
“I used to use Ollama and Whisper”
“the same model ran twice as fast”
“I use Ollama because it is easy to set up and manage”
“I've been primarily running Qwen 3.6 35B A3B out of it via Ollama”
“That's why I use Ollama.”
“I use Ollama (CLI and their npm as a dev) + HF because it just works.”
“I never hit quota limits.”
“on this VC funded joke it would be simply unusable”
“the most convenient way to run local LLMs”
“The $20/mo tier is very generous, in my experience.”
“I used ollama with Qwen models works awesome”
“I get so much usage out of it on a $20 Ollama cloud plan.”
“found a lot of performance improvement over ollama”
“Highly recommend just dropping Ollama.”
“Never reached the 5 hours usage limits of the $20 plan (on side projects)”
“The fact that I can’t easily point another tool at the same GGUF file is a dealbreaker for me.”
“Ollama is a vampire on the culture and their demise cannot come soon enough.”
“Ollama is about 1000x easier to use.”
“due to some active bugs in ollama related to reasoning and streaming”
“It runs entirely on your machine via Ollama (I recommend gemma3:27b)”