Leaderboard
Artificial Analysis Intelligence Index — a blended measure of reasoning, knowledge, and coding across open-weight and proprietary language models.
| # | Model | Intelligence | Coding | t/s | $/M |
|---|---|---|---|---|---|
| 1 | GLM-5.3 Zhipu AI | 44.9 | 74.8 | 58 | 2.15 |
| 2 | Kimi K3 Moonshot AI | 43.8 | 76.2 | 38 | 6 |
| 3 | GLM 5.3 Flash Zhipu AI | 41.9 | 71.5 | 109 | 0.24 |
| 4 | Qwen3.8 2.4T A95B Qwen · Alibaba | 40.0 | 71.9 | 39 | 3 |
| 5 | Qwen3.8-Flash-Next Qwen · Alibaba | 39.9 | 73.1 | 57 | 0.23 |
| 6 | DeepSeek V4.1 Flash DeepSeek | 39.5 | — | 221 | 0.53 |
| 7 | DeepSeek V4 Pro 0813 DeepSeek | 36.3 | 68.8 | 80 | 1.98 |
| 8 | DeepSeek V4 Flash 0731 Vision DeepSeek | 35.0 | 65 | 223 | 0.66 |
| 9 | DeepSeek V4 Flash 0731 DeepSeek | 34.5 | 69.1 | 225 | 0.66 |
| 10 | GLM-5.2 Zhipu AI | 34.0 | 68.8 | 72 | 2.15 |
| 11 | Qwen3.8 27B Qwen · Alibaba | 33.9 | 68.1 | 43 | 1.13 |
| 12 | GLM-5 Zhipu AI | 27.9 | — | — | 1.55 |
| 13 | Kimi K2.6 Moonshot AI | 27.5 | 61.8 | — | 1.71 |
| 14 | GLM-5.1 Zhipu AI | 26.4 | 55.8 | — | 2 |
| 15 | Kimi K2.7 Code Moonshot AI | 26.3 | 60.8 | 54 | 1.71 |
| 16 | Solar Open2 250B Upstage | 24.7 | 45 | — | — |
| 17 | Kimi K2.5 Moonshot AI | 23.5 | 46.8 | — | 1.14 |
| 18 | Nemotron 3 Ultra 550B A55B NVIDIA | 23.4 | 49.3 | 164 | 1.05 |
| 19 | Qwen3.5 27B Qwen · Alibaba | 22.9 | — | — | 0.83 |
| 20 | GLM-4.7 Zhipu AI | 22.2 | 45.3 | — | 1 |
Leaderboard data is cached from Artificial Analysis and refreshed every 12 hours. Scores shown are for reference; see the source for full methodology.