Public preview · figures are illustrative fixture data, attributed to their source · data to May 2025 · read-only

Models

25 open-weight models and what they need to run

Reasoning5 of 25

Models with reasoning ability, strongest reasoning benchmark results first.Show all models

Coding Run locally Reasoning Vision Agents & tool use Fast models Small models Permissive licenses

DeepSeek-R1 671B
Very large general-purpose model for coding and reasoning. Among the strongest open models at coding.
DeepSeek
- ~169 GB to run · needs server-class memory
- Runs only slowly on reference systems
- Permissive license · Free to use, including commercially
View model →
Qwen2.5 32B
General-purpose model for coding, reasoning and tool use. Among the strongest open models at reasoning.
Qwen Team (Alibaba Cloud)
- ~21 GB to run · fits a 24 GB graphics card
- Runs well on most reference systems
- Permissive license · Free to use, including commercially
View model →
Qwen2.5-Math 7B
Compact model for maths and reasoning.
Qwen Team (Alibaba Cloud)
- ~6 GB to run · fits most laptops
- Runs well on most reference systems
- Permissive license · Free to use, including commercially
View model →
Qwen3 30B-A3B
Efficient general-purpose model for coding, reasoning and tool use.
Qwen Team (Alibaba Cloud)
- ~18 GB to run · fits a 24 GB graphics card
- Runs well on most reference systems
- Permissive license · Free to use, including commercially
View model →
Phi-4 14B
General-purpose model for coding and reasoning.
Microsoft
- ~10 GB to run · fits a 16 GB graphics card
- Runs well on most reference systems
- Permissive license · Free to use, including commercially
View model →

Parameters, context length, capability profiles and side-by-side comparison are in the Technical view.