Reference table
AI model ranking.
Compare leading AI models by benchmark score, provider, API price, and output speed.
Benchmark board
Model ranking
Rank, model, provider, Artificial Analysis Intelligence Index, blended API price, and output speed.
| Rank | Model | Provider | Index | Blended price | Speed |
|---|---|---|---|---|---|
| #1 | GPT-5.5 (xhigh) | OpenAI | 60.2 | $11.250/M | 72.3 tok/s |
| #2 | GPT-5.5 (high) | OpenAI | 58.9 | $11.250/M | 66.2 tok/s |
| #3 | Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | Anthropic | 57.3 | $10.938/M | 54.3 tok/s |
| #4 | Gemini 3.1 Pro Preview | 57.2 | $4.500/M | 125.3 tok/s | |
| #5 | GPT-5.4 (xhigh) | OpenAI | 56.8 | $5.625/M | 89.7 tok/s |
| #6 | GPT-5.5 (medium) | OpenAI | 56.7 | $11.250/M | 68.1 tok/s |
| #7 | Qwen3.7 Max | Qwen | 56.6 | $3.750/M | 204.7 tok/s |
| #8 | Gemini 3.5 Flash (high) | 55.3 | $3.375/M | 230.0 tok/s | |
| #9 | Kimi K2.6 | Kimi | 53.9 | $1.712/M | 33.3 tok/s |
| #10 | MiMo-V2.5-Pro | Xiaomi | 53.8 | $1.350/M | 51.4 tok/s |
| #11 | GPT-5.3 Codex (xhigh) | OpenAI | 53.6 | $4.813/M | 78.5 tok/s |
| #12 | Grok 4.3 (high) | xAI | 53.2 | $1.563/M | 191.3 tok/s |
| #13 | Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | Anthropic | 52.9 | $10.938/M | 51.4 tok/s |
| #14 | Muse Spark | Meta | 52.2 | Not listed | Not listed |
| #15 | Claude Opus 4.7 (Non-reasoning, High Effort) | Anthropic | 51.8 | $10.938/M | 45.8 tok/s |
| #16 | Qwen3.6 Max Preview | Qwen | 51.8 | $2.925/M | 35.9 tok/s |
| #17 | Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | Anthropic | 51.7 | $6.563/M | 66.5 tok/s |
| #18 | DeepSeek V4 Pro (Reasoning, Max Effort) | DeepSeek | 51.5 | $0.544/M | 52.3 tok/s |
| #19 | GLM-5.1 (Reasoning) | Z.ai | 51.4 | $2.150/M | 59.9 tok/s |
| #20 | GPT-5.2 (xhigh) | OpenAI | 51.3 | $4.813/M | 81.2 tok/s |
| #21 | GPT-5.5 (low) | OpenAI | 50.8 | $11.250/M | 69.7 tok/s |
| #22 | Qwen3.6 Plus | Qwen | 50.0 | $1.125/M | 52.7 tok/s |
| #23 | DeepSeek V4 Pro (Reasoning, High Effort) | DeepSeek | 49.8 | $0.544/M | 47.9 tok/s |
| #24 | GLM-5 (Reasoning) | Z.ai | 49.8 | $1.550/M | 72.0 tok/s |
| #25 | Claude Opus 4.5 (Reasoning) | Anthropic | 49.7 | $10.938/M | 69.4 tok/s |
| #26 | MiniMax-M2.7 | MiniMax | 49.6 | $0.525/M | 58.5 tok/s |
| #27 | Grok 4.20 0309 v2 (Reasoning) | xAI | 49.3 | $3.000/M | 176.5 tok/s |
| #28 | MiMo-V2-Pro | Xiaomi | 49.2 | $1.500/M | 62.8 tok/s |
| #29 | GPT-5.2 Codex (xhigh) | OpenAI | 49.0 | $4.813/M | 113.3 tok/s |
| #30 | MiMo-V2.5 | Xiaomi | 49.0 | $0.408/M | 94.5 tok/s |
| #31 | GPT-5.4 mini (xhigh) | OpenAI | 48.9 | $1.688/M | 160.2 tok/s |
| #32 | Grok 4.3 (medium) | xAI | 48.8 | $1.563/M | 168.4 tok/s |
| #33 | Grok 4.20 0309 (Reasoning) | xAI | 48.5 | $3.000/M | 181.8 tok/s |
| #34 | Gemini 3 Pro Preview (high) | 48.4 | $4.500/M | 133.1 tok/s | |
| #35 | GPT-5.4 (low) | OpenAI | 47.9 | $5.625/M | 74.3 tok/s |
| #36 | GPT-5.1 (high) | OpenAI | 47.7 | $3.438/M | 137.5 tok/s |
| #37 | GLM-5-Turbo | Z.ai | 46.8 | Not listed | Not listed |
| #38 | Kimi K2.5 (Reasoning) | Kimi | 46.8 | $1.185/M | 56.7 tok/s |
| #39 | GPT-5.2 (medium) | OpenAI | 46.6 | $4.813/M | Not listed |
| #40 | Claude Opus 4.6 (Non-reasoning, High Effort) | Anthropic | 46.5 | $10.938/M | 52.2 tok/s |
| #41 | DeepSeek V4 Flash (Reasoning, Max Effort) | DeepSeek | 46.5 | $0.175/M | 113.1 tok/s |
| #42 | Gemini 3 Flash Preview (Reasoning) | 46.4 | $1.125/M | 199.1 tok/s | |
| #43 | DeepSeek V4 Flash (Reasoning, High Effort) | DeepSeek | 46.0 | $0.175/M | Not listed |
| #44 | Qwen3.6 27B (Reasoning) | Qwen | 45.8 | $1.350/M | 63.0 tok/s |
| #45 | Qwen3.5 397B A17B (Reasoning) | Qwen | 45.0 | $1.350/M | 51.9 tok/s |
| #46 | MiMo-V2-Omni-0327 | Xiaomi | 44.9 | $0.800/M | 113.9 tok/s |
| #47 | GPT-5 (high) | OpenAI | 44.6 | $3.438/M | 93.1 tok/s |
| #48 | GPT-5 Codex (high) | OpenAI | 44.6 | $3.438/M | 198.8 tok/s |
| #49 | Claude Sonnet 4.6 (Non-reasoning, High Effort) | Anthropic | 44.4 | $6.563/M | 55.2 tok/s |
| #50 | GPT-5.4 nano (xhigh) | OpenAI | 44.0 | $0.463/M | 150.7 tok/s |
| #51 | Grok 4.3 (low) | xAI | 43.9 | $1.563/M | 151.1 tok/s |
| #52 | GLM-5.1 (Non-reasoning) | Z.ai | 43.8 | $2.150/M | 45.7 tok/s |
| #53 | KAT Coder Pro V2 | KwaiKAT | 43.8 | $0.525/M | 113.7 tok/s |
| #54 | Qwen3.6 35B A3B (Reasoning) | Qwen | 43.5 | $0.557/M | 169.8 tok/s |
| #55 | MiMo-V2-Omni | Xiaomi | 43.4 | Not listed | 106.7 tok/s |
| #56 | Gemini 3.5 Flash (minimal) | 43.3 | $3.375/M | 216.9 tok/s | |
| #57 | Claude Opus 4.5 (Non-reasoning) | Anthropic | 43.1 | $10.938/M | 58.7 tok/s |
| #58 | GPT-5.1 Codex (high) | OpenAI | 43.1 | $3.438/M | 187.4 tok/s |
| #59 | Claude 4.5 Sonnet (Reasoning) | Anthropic | 43.0 | $6.563/M | 54.5 tok/s |
| #60 | GLM 5V Turbo (Reasoning) | Z.ai | 42.9 | Not listed | Not listed |
| #61 | Kimi K2.6 (Non-reasoning) | Kimi | 42.9 | $1.712/M | 46.9 tok/s |
| #62 | Claude Sonnet 4.6 (Non-reasoning, Low Effort) | Anthropic | 42.6 | $6.563/M | 51.8 tok/s |
| #63 | GLM-4.7 (Reasoning) | Z.ai | 42.1 | $1.000/M | 98.7 tok/s |
| #64 | Qwen3.5 27B (Reasoning) | Qwen | 42.1 | $0.825/M | 83.2 tok/s |
| #65 | Claude 4.1 Opus (Reasoning) | Anthropic | 42.0 | $32.813/M | 40.9 tok/s |
| #66 | GPT-5 (medium) | OpenAI | 42.0 | $3.438/M | 91.0 tok/s |
| #67 | Hy3-preview (Reasoning) | Tencent | 41.9 | $0.200/M | 93.9 tok/s |
| #68 | MiniMax-M2.5 | MiniMax | 41.9 | $0.525/M | 101.0 tok/s |
| #69 | DeepSeek V3.2 (Reasoning) | DeepSeek | 41.7 | $0.337/M | Not listed |
| #70 | Qwen3.5 122B A10B (Reasoning) | Qwen | 41.6 | $1.100/M | 146.6 tok/s |
| #71 | Grok 4 | xAI | 41.5 | $11.000/M | Not listed |
| #72 | MiMo-V2-Flash (Feb 2026) | Xiaomi | 41.5 | $0.150/M | 131.6 tok/s |
| #73 | Gemini 3 Pro Preview (low) | 41.3 | $4.500/M | Not listed | |
| #74 | GPT-5 mini (high) | OpenAI | 41.2 | $0.688/M | 90.8 tok/s |
| #75 | GPT-5.5 (Non-reasoning) | OpenAI | 40.9 | $11.250/M | 65.5 tok/s |
| #76 | Kimi K2 Thinking | Kimi | 40.9 | $1.075/M | 100.0 tok/s |
| #77 | o3-pro | OpenAI | 40.7 | $35.000/M | 25.5 tok/s |
| #78 | GLM-5 (Non-reasoning) | Z.ai | 40.6 | $1.550/M | 64.0 tok/s |
| #79 | Qwen3.5 397B A17B (Non-reasoning) | Qwen | 40.1 | $1.350/M | 53.0 tok/s |
| #80 | Qwen3 Max Thinking | Qwen | 39.8 | $2.400/M | 50.6 tok/s |
Sources
Updated 2026-05-27. Benchmark methodology and ranking source below.