Ollama
A free app that runs AI models on your own computer. It's for people who write code. It's free.
Get patch notes
Same daily newsletter as the rest of the site. One email a day.
No affiliate link on file. Outbound goes to ollama.com. Identity checked October 21, 2026. Score null is from the run on October 21, 2026.
Grade
We tried it ourselves. Grade B.
- Grade
- B · We tried it
- Score
- 7/10
- Price
- Free
Best for
Jobs people use this for.
Test details
Runs open-weight LLMs locally behind a CLI and a REST API. MIT license.
Hands-on review · 2026-10-21
TestedB7/10- CPU-only box: 8 cores, 15 GB RAM, no GPU.
- Registry pull was not tested because the sandbox network blocked it
Ollama itself is easy: once zstd was in place it installed in 17 seconds, and it serves models at usable CPU speeds. The small models that fit a CPU-only box are the weak point. All three wrote correct code, none got either trick question right, and none kept the summary under 60 words. We had to import models from Hugging Face files because the registry download was blocked on our network, so this is not a clean test of `ollama pull`.
Runtime installs fast and runs well on CPU. Scored on the runtime, not the small models' poor reasoning.
- Coverage
- Install, plus the 4-task battery on three small CPU models
- What we tested
- Install friction, plus the same 4-task battery on three small CPU models: Qwen2.5 0.5B, Llama 3.2 1B and Qwen2.5-Coder 1.5B (Q4_K_M GGUF).
Method
Official install.sh, Ollama 0.40.7. Models were imported from Hugging Face GGUF files with `ollama create` because `ollama pull` failed on this box's network. Temperature 0, seed 1, num_ctx 4096.
Measured
- installFirst try failed (install.sh needs zstd). Worked in 17 s after apt install zstd
- pull`ollama pull` failed: the blob host resolved to a blocked address on this sandbox's network. Probably an environment issue, not an Ollama bug
- qwen2.5-0.5b.codepassed, 61.8 tok/s, wall 3.05 s
- qwen2.5-0.5b.reason1failed, 95.6 tok/s, wall 0.17 s
- qwen2.5-0.5b.reason2failed, 49.6 tok/s, wall 0.25 s
- qwen2.5-0.5b.summaryfailed, 58.6 tok/s, wall 7.16 s
- llama3.2-1b.codepassed, 30.3 tok/s, wall 4.0 s
- llama3.2-1b.reason1failed, 49.2 tok/s, wall 0.2 s
- llama3.2-1b.reason2failed, 39.2 tok/s, wall 0.46 s
- llama3.2-1b.summaryfailed, 29.2 tok/s, wall 5.49 s
- qwen2.5-coder-1.5b.codepassed, 28.9 tok/s, wall 5.26 s
- qwen2.5-coder-1.5b.reason1failed, 36.7 tok/s, wall 0.24 s
- qwen2.5-coder-1.5b.reason2failed, 35.1 tok/s, wall 0.27 s
- qwen2.5-coder-1.5b.summaryfailed, 23.0 tok/s, wall 8.33 s
- summaryCode 3/3 models passed. Reasoning 0/6 (answers: '10', '$1.05', '500 minutes'). Summary 0/3 kept to 60 words. 23-96 tokens/s on CPU
Pros
- One-line install, fast local API
- Usable speed on CPU for 0.5-1.5B models (23-62 tok/s on longer outputs)
- Fully offline once models are on disk
Cons
- install.sh needs zstd and doesn't install it
- Small models failed every trick question and ignored the length limit
- Models imported from GGUF (our workaround) had no tool-calling template, which broke tool use in LangChain and Open Interpreter
- Status
- Not run
- Date
- October 21, 2026
- Time to first output
- —
- Cost
- —
- Result
- —
- Evidence
- Log linked below
Queued task. Answer a fixed research brief. Log whether claims cite sources, time to a usable answer, and cost.
FAQ
What is the agent-tested verdict for Ollama?
A free app that runs AI models on your own computer. It's for people who write code. It's free. We tried it ourselves. Grade B.
How much does Ollama cost?
Free. MIT. Local open-weight runtime.
Do sponsors outrank this tool?
No. A sponsored slot is labeled and sold separately from tiers. It cannot move this listing.