Side-by-side benchmarks, capability radar charts, and use-case rankings across Gemma 4, GPT-4o, Claude 3.5 Sonnet, Llama 4, Qwen 3.5, and Mistral Small 4.
Normalised scores across six capability dimensions. Larger area = stronger overall profile. Toggle models above to add / remove layers.
| Model | Developer | Params (Total) | Active Params | Context | Architecture | License | Vision | Audio | Thinking | Fn Call | Min VRAM | Arena ELO | MMLU Pro |
|---|