AI Model Leaderboard
The only board that ranks LLMs on all three axes that matter to a builder: how smart they are (intelligence), how widely used they are (real OpenRouter traffic), and your value — intelligence per dollar.
Updated July 20, 2026
Ranked by the Artificial Analysis Intelligence Index. Value = intelligence per blended $/1M token. Popularity shown where the model has OpenRouter traffic.
| # | Model | |||
|---|---|---|---|---|
| 1 | Claude Fable 5 | 59.9 | #31 4% | 2.00 |
| 2 | GPT-5.6 Sol | 58.9 | #42 97% | 3.37 |
| 3 | Kimi K3 | 57.1 | #49 — | 6.34 |
| 4 | Claude Opus 4.8 | 55.7 | #8 22% | 3.71 |
| 5 | GPT-5.6 Terra | 55.0 | #55 436% | 6.29 |
| 6 | GPT-5.5 | 54.8 | #14 36% | 3.13 |
| 7 | Grok 4.5 | 53.8 | #47 15% | 13.45 |
| 8 | Claude Opus 4.7 | 53.5 | #7 15% | 3.57 |
| 9 | Claude Sonnet 5 | 53.4 | #16 2% | 8.90 |
| 10 | GPT-5.4 | 51.4 | #24 34% | 5.87 |
| 11 | GPT-5.6 Luna | 51.2 | #50 159% | 14.63 |
| 12 | GLM-5.2 | 51.1 | #5 0% | 17.62 |
| 13 | Muse Spark 1.1 | 50.6 | — | 18.40 |
| 14 | Gemini 3.5 Flash | 50.2 | #23 20% | 9.56 |
| 15 | Claude Sonnet 4.6 | 47.2 | #12 20% | 5.24 |
| 16 | Gemini 3.1 Pro Preview | 46.5 | #35 22% | 6.64 |
| 17 | Qwen3.7 Max | 46.0 | #37 17% | 9.20 |
| 18 | MiniMax-M3 | 44.4 | #4 25% | 59.20 |
| 19 | GPT-5.3 Codex | 44.3 | — | 5.63 |
| 20 | DeepSeek V4 Pro | 44.3 | #6 15% | 68.15 |
| 21 | Kimi K2.6 | 44.2 | #27 27% | 17.86 |
| 22 | Claude Opus 4.6 | 43.7 | #33 15% | 2.91 |
| 23 | Muse Spark | 43.1 | — | — |
| 24 | MiMo-V2.5-Pro | 42.2 | #19 21% | 64.92 |
| 25 | GPT-5.2 | 42.2 | #59 — | 5.36 |
| 26 | Kimi K2.7 Code | 41.9 | #38 30% | 16.93 |
| 27 | Hy3 | 41.2 | #3 34% | 4120.00 |
| 28 | Nex-N2-Pro | 41.0 | #62 — | 27.33 |
| 29 | Claude Opus 4.5 | 40.8 | — | 2.72 |
| 30 | Inkling | 40.7 | — | 12.43 |
| 31 | MiMo-V2-Pro | 40.3 | — | — |
| 32 | DeepSeek V4 Flash | 40.3 | #2 12% | 191.90 |
| 33 | GLM-5.1 | 40.2 | #41 44% | 13.86 |
| 34 | GPT-5.2 Codex | 40.1 | — | 5.09 |
| 35 | GPT-5.4 mini | 40.0 | #36 39% | 15.24 |
| 36 | Qwen3.6 Max Preview | 40.0 | — | 8.79 |
| 37 | Grok Build 0.1 0616 | 39.8 | — | 26.53 |
| 38 | Gemini 3 Pro Preview | 39.6 | — | 5.66 |
| 39 | Qwen3.6 Plus | 39.6 | — | 22.63 |
| 40 | GLM-5 | 39.5 | #34 63% | 18.81 |
| 41 | Qwen3.7 Plus | 39.0 | #40 40% | 39.00 |
| 42 | JT-4.1 Flash 236B A21B | 38.8 | — | — |
| 43 | GPT-5.4 nano | 38.2 | #44 2% | 52.69 |
| 44 | GLM-5-Turbo | 38.1 | — | — |
| 45 | MiniMax-M2.7 | 38.1 | #46 65% | 50.80 |
| 46 | Nemotron 3 Ultra 550B A55B | 37.8 | #10 8% | 22.57 |
| 47 | Gemini 3 Flash Preview | 37.8 | #13 14% | 21.60 |
| 48 | Grok 4.3 | 37.6 | #52 100% | 20.05 |
| 49 | MiMo-V2.5 | 37.2 | #1 21% | 177.14 |
| 50 | Qwen3.6 27B | 37.1 | — | 17.67 |
The board is a superset of the calculator catalog: brand-new models appear here with quality & popularity before the pricing catalog catches up. Models with a calculator match link to a head-to-head comparison.
Data: Artificial Analysis · OpenRouter
Know your margin before you build
Pick a model above, then model your real cost and price in the Calcaas calculator — free.