Best Local OpenAI Models
The best local OpenAI model is gpt-oss-120b (23.8 intelligence) — open weights on Hugging Face, small enough to run on your own hardware. gpt-oss-20b (14.9) and gpt-oss-20b (free) (14.9) round out the top three.
The best local LLMs — open-weight models small enough to run on your own hardware with Ollama, llama.cpp or vLLM — ranked by intelligence. Frontier-scale open models (400B+) are excluded: downloadable isn't the same as runnable. Every model here is on Hugging Face.
- 1
gpt-oss-120bReasoningToolsJSON23.8 intel · $0.037/M · 131K ctx23.8Intelligence - 2gpt-oss-20bReasoningToolsJSON14.9 intel · $0.029/M · 131K ctx14.9Intelligence
- 3gpt-oss-20b:freeReasoningToolsJSON14.9 intel · Free/M · 131K ctx14.9Intelligence
Frequently asked
What is the best local OpenAI model?
The best local OpenAI model is gpt-oss-120b (23.8 intelligence) — open weights on Hugging Face, small enough to run on your own hardware. gpt-oss-20b (14.9) and gpt-oss-20b (free) (14.9) round out the top three.
What's a good alternative to gpt-oss-120b?
gpt-oss-20b (14.9) is the closest alternative on this metric, followed by gpt-oss-20b (free) (14.9). See the full ranking above for the tradeoffs.
How many OpenAI models are there?
modelgrep tracks 67 OpenAI models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by GPT-5.6 Sol. 3 of them qualify for this ranking.