Reasoning leaderboard across GPQA, HLE, MMLU-Pro, and related metrics.
Domain score averages relative percentile across included metrics.
Domain score 100
Reasoning leaderboard across GPQA, HLE, MMLU-Pro, and related metrics.
| Rank | Model | Creator | Domain Score | Speed | Blended Price |
|---|---|---|---|---|---|
| #1 | Gemini 3.1 Pro Preview | 99.9 | 135.9 tok/s | $4.50/M | |
| #2 |
| OpenAI |
| 99.6 |
| 62.4 tok/s |
| $11.25/M |
| #3 | GPT-5.5 (high) | OpenAI | 99.4 | 62.3 tok/s | $11.25/M |
| #4 | GPT-5.4 (xhigh) | OpenAI | 99.1 | 75.5 tok/s | $5.63/M |
| #5 | Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | Anthropic | 99.0 | 66.3 tok/s | $10.94/M |
| #6 | GPT-5.5 (medium) | OpenAI | 98.8 | 59.2 tok/s | $11.25/M |
| #7 | Gemini 3 Pro Preview (high) | 98.5 | n/a | $4.50/M |
| #8 | Gemini 3.5 Flash (high) | 98.5 | 212.4 tok/s | $3.38/M |
| #9 | Gemini 3.5 Flash (medium) | 98.3 | 210.2 tok/s | $3.38/M |
| #10 | Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | Anthropic | 98.2 | 61.8 tok/s | $10.94/M |
| #11 | GPT-5.3 Codex (xhigh) | OpenAI | 98.1 | 98.6 tok/s | $4.81/M |
| #12 | Kimi K2.6 | Kimi | 97.5 | 41.6 tok/s | $1.71/M |
| #13 | GPT-5.2 (xhigh) | OpenAI | 97.0 | 71 tok/s | $4.81/M |
| #14 | Qwen3.7 Max | Alibaba | 97.0 | 186.5 tok/s | $3.75/M |
| #15 | Gemini 3 Flash Preview (Reasoning) | 96.9 | 172.8 tok/s | $1.13/M |
| #16 | GPT-5.2 Codex (xhigh) | OpenAI | 96.9 | 105.3 tok/s | $4.81/M |
| #17 | Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | Anthropic | 96.6 | 47.3 tok/s | $10.94/M |
| #18 | Muse Spark | Meta | 96.2 | n/a | - |
| #19 | GPT-5.5 (low) | OpenAI | 96.2 | 60.2 tok/s | $11.25/M |
| #20 | DeepSeek V4 Pro (Reasoning, Max Effort) | DeepSeek | 95.8 | 61.8 tok/s | $0.544/M |
| #21 | MiniMax-M3 | MiniMax | 95.8 | 45.6 tok/s | $0.525/M |
| #22 | Grok 4.3 (high) | xAI | 95.5 | 237.5 tok/s | $1.56/M |
| #23 | Gemini 3 Pro Preview (low) | 95.2 | n/a | $4.50/M |
| #24 | Claude Opus 4.7 (Non-reasoning, High Effort) | Anthropic | 94.8 | 48.1 tok/s | $10.94/M |
| #25 | DeepSeek V4 Pro (Reasoning, High Effort) | DeepSeek | 94.8 | 57.6 tok/s | $0.544/M |
| #26 | Claude Opus 4.5 (Reasoning) | Anthropic | 94.6 | 53.5 tok/s | $10.94/M |
| #27 | Grok 4.20 0309 v2 (Reasoning) | xAI | 94.5 | 168.7 tok/s | $3.00/M |
| #28 | MiMo-V2.5-Pro | Xiaomi | 94.2 | 43.3 tok/s | $0.544/M |
| #29 | Qwen3.7 Plus | Alibaba | 94.0 | 53.6 tok/s | $0.590/M |
| #30 | GPT-5.4 (low) | OpenAI | 94.0 | 63.6 tok/s | $5.63/M |
| #31 | Kimi K2.5 (Reasoning) | Kimi | 93.9 | 31.7 tok/s | $1.19/M |
| #32 | Qwen3.6 Max Preview | Alibaba | 93.5 | 40.9 tok/s | $2.93/M |
| #33 | DeepSeek V4 Flash (Reasoning, Max Effort) | DeepSeek | 93.4 | 107.8 tok/s | $0.175/M |
| #34 | Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | Anthropic | 93.2 | 70.4 tok/s | $6.56/M |
| #35 | GPT-5.4 mini (xhigh) | OpenAI | 93.0 | 176 tok/s | $1.69/M |
| #36 | MiniMax-M2.7 | MiniMax | 92.8 | 75 tok/s | $0.525/M |
| #37 | Grok 4.20 0309 (Reasoning) | xAI | 92.5 | 166.5 tok/s | $3.00/M |
| #38 | Grok 4.3 (medium) | xAI | 92.4 | 197.6 tok/s | $1.56/M |
| #39 | Grok 4 | xAI | 92.1 | n/a | $11.00/M |
| #40 | GPT-5.1 (high) | OpenAI | 91.8 | 121.2 tok/s | $3.44/M |
| #41 | DeepSeek V3.2 Speciale | DeepSeek | 91.6 | n/a | - |
| #42 | GPT-5.2 (medium) | OpenAI | 91.0 | n/a | $4.81/M |
| #43 | GPT-5 (high) | OpenAI | 90.7 | 111.1 tok/s | $3.44/M |
| #44 | GLM-5.1 (Reasoning) | Z AI | 90.7 | 46.8 tok/s | $2.15/M |
| #45 | GLM-4.7 (Reasoning) | Z AI | 90.4 | 79.2 tok/s | $1.00/M |
| #46 | Qwen3.5 397B A17B (Reasoning) | Alibaba | 90.3 | 51.8 tok/s | $1.35/M |
| #47 | MiMo-V2-Pro | Xiaomi | 90.1 | 42.5 tok/s | $1.50/M |
| #48 | GPT-5.5 Instant (May 2026) | OpenAI | 89.4 | n/a | $11.25/M |
| #49 | DeepSeek V4 Flash (Reasoning, High Effort) | DeepSeek | 89.3 | n/a | $0.175/M |
| #50 | Qwen3 Max Thinking | Alibaba | 89.3 | n/a | $2.40/M |
| #51 | GLM-5-Turbo | Z AI | 88.3 | n/a | - |
| #52 | GPT-5 (medium) | OpenAI | 88.2 | 85.6 tok/s | $3.44/M |
| #53 | Gemini 2.5 Pro | 88.2 | 139.8 tok/s | $3.44/M |
| #54 | Qwen3.6 Plus | Alibaba | 88.2 | 52.8 tok/s | $1.13/M |
| #55 | Gemini 3.5 Flash (minimal) | 88.1 | 199.1 tok/s | $3.38/M |
| #56 | Claude 4.5 Sonnet (Reasoning) | Anthropic | 88.1 | 50.1 tok/s | $6.56/M |
| #57 | Gemma 4 31B (Reasoning) | 88.1 | 34.8 tok/s | - |
| #58 | MiMo-V2.5 | Xiaomi | 88.1 | 77.4 tok/s | $0.175/M |
| #59 | Hy3-preview (Reasoning) | Tencent | 88.0 | 96 tok/s | $0.200/M |
| #60 | GLM-5 (Reasoning) | Z AI | 87.9 | 79.5 tok/s | $1.55/M |
| #61 | GPT-5.4 nano (xhigh) | OpenAI | 87.9 | 158.3 tok/s | $0.463/M |
| #62 | GPT-5 Codex (high) | OpenAI | 87.7 | 171.1 tok/s | $3.44/M |
| #63 | Grok 4.1 Fast (Reasoning) | xAI | 87.7 | n/a | - |
| #64 | Gemini 3 Flash Preview (Non-reasoning) | 87.6 | 181.3 tok/s | $1.13/M |
| #65 | GPT-5.1 Codex (high) | OpenAI | 87.6 | 182.1 tok/s | $3.44/M |
| #66 | Qwen3.5 122B A10B (Reasoning) | Alibaba | 87.0 | 143.6 tok/s | $1.10/M |
| #67 | MiniMax-M2.1 | MiniMax | 87.0 | 184.6 tok/s | $0.525/M |
| #68 | Grok 4 Fast (Reasoning) | xAI | 86.8 | n/a | $0.275/M |
| #69 | o3-pro | OpenAI | 86.6 | 22.7 tok/s | $35.00/M |
| #70 | Kimi K2 Thinking | Kimi | 86.5 | 131.1 tok/s | $1.08/M |
| #71 | Claude Opus 4.5 (Non-reasoning) | Anthropic | 86.4 | 47.6 tok/s | $10.94/M |
| #72 | Claude Opus 4.6 (Non-reasoning, High Effort) | Anthropic | 86.3 | 40.9 tok/s | $10.94/M |
| #73 | MiniMax-M2.5 | MiniMax | 85.5 | 202.9 tok/s | $0.525/M |
| #74 | Qwen3.5 397B A17B (Non-reasoning) | Alibaba | 85.2 | 53.1 tok/s | $1.35/M |
| #75 | Ring-2.6-1T | InclusionAI | 85.2 | 122.1 tok/s | $0.850/M |
| #76 | o3 | OpenAI | 85.1 | 159.6 tok/s | $3.50/M |
| #77 | DeepSeek V3.2 (Reasoning) | DeepSeek | 84.9 | n/a | $0.337/M |
| #78 | MiMo-V2-Flash (Reasoning) | Xiaomi | 84.1 | 129.5 tok/s | $0.150/M |
| #79 | Qwen3.5 27B (Reasoning) | Alibaba | 83.7 | 82.8 tok/s | $0.825/M |
| #80 | GPT-5.4 mini (medium) | OpenAI | 83.6 | 184.4 tok/s | $1.69/M |
