Best AI models for reasoning
GPQA Diamond is a set of graduate-level science questions written by domain experts and filtered so that PhD students with internet access still struggle. It's the most reliable signal we have for "does this model actually reason" vs "is it pattern-matching training data".
- 馃
gpt-6-astraOpenAI路$10.00 / $50.00 per 1M96.1%96.1% - 馃
gemini-3.8-flash50% offGoogle Vertex AI路$0.75 / $3.75 per 1M95.3%95.3% - 馃grok-4.6xAI路$2.00 / $6.00 per 1M94.9%94.9%
- 4gpt-5.6-solOpenAI路$4.00 / $20.00 per 1M94.1%94.1%
- 5
gemini-3.1-pro-previewGoogle Gemini API路$2.00 / $12.00 per 1M94.1%94.1% - 6
claude-fable-5.1Anthropic路$10.00 / $50.00 per 1M93.7%93.7% - 7muse-spark-1.3meta路$1.25 / $4.25 per 1M93.5%93.5%
- 8qwen3.8-2.4t-a95bTensorX路$2.50 / $6.00 per 1M93.5%93.5%
- 9
kimi-k3Moonshot AI路$3.00 / $15.00 per 1M93.5%93.5% - 10gpt-5.5OpenAI路$5.00 / $30.00 per 1M93.5%93.5%
- 11claude-opus-5Anthropic路$5.00 / $25.00 per 1M93.2%93.2%
- 12grok-4.5xAI路$2.00 / $6.00 per 1M93.1%93.1%
- 13minimax-m350% offMiniMax路$0.30 / $1.20 per 1M92.9%92.9%
- 14
deepseek-v4-pro-0813DeepSeek路$1.32 / $3.96 per 1M92.8%92.8% - 15
qwen3.8-maxAlibaba Cloud路$2.00 / $6.00 per 1M92.8%92.8% - 16gemini-3.6-flashGoogle Gemini API路$1.50 / $7.00 per 1M92.8%92.8%
- 17claude-fable-5Anthropic路$10.00 / $50.00 per 1M92.6%92.6%
- 18gpt-5.6-terraOpenAI路$2.00 / $12.00 per 1M92.5%92.5%
- 19qwen3.8-flash-nextTensorX路$0.20 / $0.50 per 1M92.3%92.3%
- 20qwen3.7-maxAlibaba Cloud路$2.50 / $7.50 per 1M92.3%92.3%
- 21gemini-3.7-flash50% offGoogle Vertex AI路$0.75 / $3.75 per 1M92.1%92.1%
- 22gemini-3.5-flashGoogle Gemini API路$1.50 / $9.00 per 1M92.1%92.1%
- 23claude-opus-4-8Anthropic路$5.00 / $25.00 per 1M92.0%92.0%
- 24gpt-5.4OpenAI路$2.50 / $15.00 per 1M92.0%92.0%
- 25
glm-5.3Z.ai路$1.40 / $4.40 per 1M91.7%91.7% - 26gpt-5.3-codexOpenAI路$1.75 / $14.00 per 1M91.5%91.5%
- 27claude-opus-4-7Anthropic路$5.00 / $25.00 per 1M91.4%91.4%
- 28glm-5.3-flashZ.ai路$0.15 / $0.50 per 1M91.2%91.2%
- 29gpt-5.6-lunaOpenAI路$0.20 / $1.20 per 1M91.1%91.1%
- 30claude-sonnet-5Anthropic路$2.00 / $10.00 per 1M91.1%91.1%
Explore other rankings
How we rank
Scores for GPQA Diamond come from Artificial Analysis, an independent AI benchmarking service. When a model is available through multiple providers (e.g. Anthropic direct, AWS Bedrock, Google Vertex), we show one canonical entry per model family so the ranking isn't polluted by duplicates. Benchmarks measure specific skills, so always validate on your own workload before committing.
One API for every model on this list
Requesty is OpenAI-compatible and routes to 600+ models. Switch between any of the models above by changing one parameter in your code.