Extended Reasoning
Whether the model supports extended/explicit reasoning modes.
Higher is better · Weight in overall score: 0.075
Top models by Extended Reasoning
| # | Name | Value | Confidence |
|---|---|---|---|
| 1 | Claude Opus 4.8Anthropic | Yes | Verified · 2+ sources |
| 2 | Claude Sonnet 4.6Anthropic | Yes | Verified · 2+ sources |
| 3 | Claude Haiku 4.5 (latest)Anthropic | Yes | Verified · 2+ sources |
| 4 | GPT-5.5OpenAI | Yes | Verified · 2+ sources |
| 5 | o4-miniOpenAI | Yes | Verified · 2+ sources |
| 6 | Gemini 3.1 Pro PreviewGoogle | Yes | Verified · 2+ sources |
| 7 | Gemini 3.5 FlashGoogle | Yes | Verified · 2+ sources |
| 8 | DeepSeek V4 ProDeepSeek | Yes | Verified · 2+ sources |
| 9 | Grok 4.20 (Reasoning)xAI | Yes | Verified · 2+ sources |
| 10 | Qwen3.7 MaxAlibaba | Yes | Verified · 2+ sources |
| 11 | Kimi K2.6Moonshot AI | Yes | Verified · 2+ sources |
| 12 | GPT-5.5 ProOpenAI | Yes | Verified · 2+ sources |
| 13 | DeepSeek V4 FlashDeepSeek | Yes | Verified · 2+ sources |
| 14 | Devstral 2Mistral | No | Verified · 2+ sources |
| 15 | Llama 4 Maverick 17B InstructMeta | No | Verified · 2+ sources |
| 16 | Mistral Large 3Mistral | No | Verified · 2+ sources |
| 17 | Qwen3 Coder PlusAlibaba | No | Verified · 2+ sources |