# Best Chinese AI Models (2026)

> A live ranking of AI models from Chinese labs, using the same current public ranking contract as the overall leaderboard.

Use this page for the current Chinese-model order. The table rebuilds from the active public ranking lane rather than preserving a copied winner, so the first row changes with the same contract used by the overall leaderboard.

Canonical page: https://benchlm.ai/best/chinese-models

Last updated: September 4, 2026

## Rankings

| Rank | Model | Creator | Type | Context | Score | Evidence | 90% interval |
|------|-------|---------|------|---------|-------|----------|--------------|
| 1 | [Kimi K3](/models/kimi-k3) | Moonshot AI | Pending | 1.05M | 74.9 | Supported | 71.40–78.35 |
| 2 | [Qwen3.8 Max](/models/qwen3-8-max) | Alibaba | Open Weight | 1M | 72.4 | Supported | 68.68–76.18 |
| 3 | [Qwen3.7 Max](/models/qwen3-7-max) | Alibaba | Proprietary | 1M | 68.6 | Supported | 62.50–74.61 |
| 4 | [GLM-5.3](/models/glm-5-3) | Z.AI | Open Weight | 1M | 68.5 | Estimated | 61.36–75.57 |
| 5 | [Qwen3.8-27B](/models/qwen3-8-27b) | Alibaba | Open Weight | 262K | 68.4 | Supported | 66.09–70.61 |
| 6 | [GLM-5.2](/models/glm-5-2) | Z.AI | Open Weight | 1M | 68.2 | Supported | 61.75–74.68 |
| 7 | [DeepSeek V4 Pro 0813](/models/deepseek-v4-pro-0813) | DeepSeek | Proprietary | 1M | 66.4 | Estimated | 54.83–77.86 |
| 8 | [GLM-5.3-Flash](/models/glm-5-3-flash) | Z.AI | Open Weight | 1M | 66.1 | Supported | 57.12–75.06 |
| 9 | [Hy4 preview](/models/hy4-preview) | Tencent | Open Weight | 1M | 65.9 | Estimated | 39.69–82.39 |
| 10 | [Hy3](/models/hy3) | Tencent | Open Weight | 256K | 65.8 | Supported | 57.37–74.31 |
| 11 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | Xiaomi | Proprietary | 1M | 65.4 | Supported | 56.82–73.99 |
| 12 | [Kimi K2.7 Code](/models/kimi-k2-7-code) | Moonshot AI | Open Weight | 256K | 65.4 | Estimated | 53.53–71.16 |
| 13 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | Open Weight | 256K | 65.3 | Supported | 56.90–73.70 |
| 14 | [MiMo-V2-Pro](/models/mimo-v2-pro) | Xiaomi | Proprietary | 1M | 64.9 | Supported | 56.47–73.32 |
| 15 | [GLM-5.1](/models/glm-5-1) | Z.AI | Open Weight | 203K | 64.4 | Supported | 55.56–73.26 |
| 16 | [GLM-5-Turbo](/models/glm-5-turbo) | Z.AI | Proprietary | 200K | 64.1 | Supported | 53.71–74.57 |
| 17 | [MiniMax M3](/models/minimax-m3) | MiniMax | Open Weight | 1M | 63.9 | Supported | 56.50–71.32 |
| 18 | [Qwen 3.6 Max (preview)](/models/qwen3-6-max-preview) | Alibaba | Proprietary | 256K | 63.6 | Supported | 53.57–73.71 |
| 19 | [Qwen3.7 Plus](/models/qwen3-7-plus) | Alibaba | Proprietary | 1M | 62.3 | Supported | 52.40–72.19 |
| 20 | [GLM-5](/models/glm-5) | Z.AI | Open Weight | 200K | 62.2 | Supported | 51.47–72.84 |
| 21 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | Proprietary | 1M | 61.5 | Supported | 53.04–69.94 |
| 22 | [MiMo-V2.5](/models/mimo-v2-5) | Xiaomi | Proprietary | 1M | 60.7 | Estimated | 47.80–72.23 |
| 23 | [MiMo-V2-Omni](/models/mimo-v2-omni) | Xiaomi | Proprietary | 262K | 60.6 | Supported | 50.26–70.97 |
| 24 | [GLM-5V-Turbo](/models/glm-5v-turbo) | Z.AI | Proprietary | 200K | 60.6 | Supported | 49.63–71.48 |
| 25 | [Qwen3.8-Flash-Next](/models/qwen3-8-flash-next) | Alibaba | Open Weight | 262K | 59.4 | Estimated | 47.91–70.94 |
| 26 | [GLM-5 (Reasoning)](/models/glm-5-reasoning) | Z.AI | Open Weight | 200K | 59.2 | Estimated | 47.65–70.67 |
| 27 | [Qwen3.5-122B-A10B](/models/qwen3-5-122b-a10b) | Alibaba | Open Weight | 262K | 59 | Supported | 50.15–67.86 |
| 28 | [GLM-4.7](/models/glm-4-7) | Z.AI | Open Weight | 200K | 59 | Supported | 47.31–70.67 |
| 29 | [MiniMax M2.7](/models/minimax-m2-7) | MiniMax | Open Weight | 200K | 59 | Supported | 51.62–66.34 |
| 30 | [Qwen3.5 397B (Reasoning)](/models/qwen3-5-397b-reasoning) | Alibaba | Open Weight | 128K | 58.9 | Estimated | 47.37–70.40 |
| 31 | [Kimi K2.5 (Reasoning)](/models/kimi-k2-5-reasoning) | Moonshot AI | Proprietary | 128K | 58.7 | Estimated | 47.22–70.25 |
| 32 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | Open Weight | 262K | 58.3 | Supported | 48.93–67.70 |
| 33 | [DeepSeek V3.2](/models/deepseek-v3-2) | DeepSeek | Open Weight | 128K | 57.6 | Supported | 46.51–68.73 |
| 34 | [DeepSeek V3.2 (Thinking)](/models/deepseek-v3-2-thinking) | DeepSeek | Open Weight | 128K | 57.5 | Estimated | 46.03–69.05 |
| 35 | [Qwen3 235B 2507 (Reasoning)](/models/qwen3-235b-2507-reasoning) | Alibaba | Open Weight | 128K | 57.4 | Estimated | 45.91–68.94 |
| 36 | [MiniMax M2.5](/models/minimax-m2-5) | MiniMax | Proprietary | 128K | 57.3 | Supported | 50.22–64.40 |
| 37 | [Hy3 Preview](/models/hy3-preview) | Tencent | Open Weight | 256K | 57.1 | Estimated | 45.57–68.59 |
| 38 | [GLM-4.5](/models/glm-4-5) | Z.AI | Proprietary | 128K | 57 | Estimated | 45.48–68.51 |
| 39 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | Open Weight | 128K | 56.4 | Estimated | 44.92–67.95 |
| 40 | [Qwen3.5 Flash](/models/qwen3-5-flash) | Alibaba | Proprietary | 1M | 56 | Supported | 46.10–65.95 |
| 41 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | Open Weight | 256K | 55.6 | Supported | 50.21–61.04 |
| 42 | [Qwen3 235B 2507](/models/qwen3-235b-2507) | Alibaba | Open Weight | 128K | 55.5 | Estimated | 43.95–66.98 |
| 43 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | Open Weight | 262K | 54.9 | Supported | 45.14–64.58 |
| 44 | [DeepSeek LLM 2.0](/models/deepseek-llm-2-0) | DeepSeek | Open Weight | 128K | 54 | Estimated | 42.49–65.52 |
| 45 | [GLM-4.6](/models/glm-4-6) | Z.AI | Open Weight | 200K | 53.9 | Supported | 38.93–68.95 |
| 46 | [Step 3.7 Flash](/models/step-3-7-flash) | StepFun | Open Weight | 256K | 53.2 | Estimated | 40.41–64.70 |
| 47 | [Step 3.5 Flash](/models/step-3-5-flash) | StepFun | Open Weight | 256K | 53.1 | Supported | 39.81–66.37 |
| 48 | [Qwen3.6-27B](/models/qwen3-6-27b) | Alibaba | Open Weight | 262K | 52.7 | Estimated | 46.97–58.48 |
| 49 | [DeepSeek V3.1](/models/deepseek-v3-1) | DeepSeek | Open Weight | 128K | 52.7 | Supported | 37.10–68.23 |
| 50 | [Ling 3.0 Flash](/models/ling-3-0-flash) | InclusionAI | Open Weight | 262K | 52.2 | Estimated | 40.68–63.71 |
| 51 | [MiMo-V2-Flash](/models/mimo-v2-flash) | Xiaomi | Open Weight | 256K | 52.2 | Supported | 38.10–66.26 |
| 52 | [DeepSeek V3.1 (Reasoning)](/models/deepseek-v3-1-reasoning) | DeepSeek | Open Weight | 128K | 51.8 | Supported | 33.16–70.35 |
| 53 | [Qwen3.5 Plus](/models/qwen3-5-plus) | Alibaba | Proprietary | 1M | 51.6 | Estimated | 40.12–63.15 |
| 54 | [DeepSeek-R1](/models/deepseek-r1) | DeepSeek | Open Weight | 128K | 51.3 | Supported | 37.45–65.20 |
| 55 | [Qwen3.7 Flash](/models/qwen3-7-flash) | Alibaba | Proprietary | 1M | 50.8 | Estimated | 39.24–62.27 |
| 56 | [DeepSeek Coder 2.0](/models/deepseek-coder-2-0) | DeepSeek | Open Weight | 128K | 49.8 | Estimated | 38.32–61.35 |
| 57 | [Seed 1.6](/models/seed-1-6) | ByteDance | Proprietary | 256K | 49.8 | Estimated | 38.27–61.30 |
| 58 | [DeepSeekMath V2](/models/deepseekmath-v2) | DeepSeek | Open Weight | 128K | 49.5 | Estimated | 37.94–60.96 |
| 59 | [Qwen2.5-1M](/models/qwen2-5-1m) | Alibaba | Open Weight | 1M | 49.5 | Estimated | 37.94–60.96 |
| 60 | [GLM-4.7-Flash](/models/glm-4-7-flash) | Z.AI | Open Weight | 200K | 49.4 | Supported | 36.37–62.44 |
| 61 | [Seed-2.0-Lite](/models/seed-2-0-lite) | ByteDance | Proprietary | 256K | 49.4 | Estimated | 37.89–60.92 |
| 62 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | Alibaba | Open Weight | 262K | 48.3 | Estimated | 42.58–54.09 |
| 63 | [GLM-4.5-Air](/models/glm-4-5-air) | Z.AI | Proprietary | 128K | 46.3 | Supported | 28.68–63.98 |
| 64 | [Seed 1.6 Flash](/models/seed-1-6-flash) | ByteDance | Proprietary | 256K | 44.8 | Estimated | 33.24–56.27 |
| 65 | [Moonshot v1](/models/moonshot-v1) | Moonshot AI | Proprietary | 128K | 44.5 | Estimated | 32.96–55.98 |
| 66 | [Seed-2.0-Mini](/models/seed-2-0-mini) | ByteDance | Proprietary | 256K | 44.3 | Estimated | 32.77–55.79 |
| 67 | [Qwen3 Max](/models/qwen3-max) | Alibaba | Proprietary | 1M | 43.8 | Estimated | 38.06–49.58 |
| 68 | [DeepSeek V3](/models/deepseek-v3) | DeepSeek | Open Weight | 128K | 43.7 | Supported | 27.74–59.73 |
| 69 | [Ling 2.6 Flash](/models/ling-2-6-flash) | InclusionAI | Open Weight | 262K | 43.7 | Estimated | 32.16–55.19 |
| 70 | [Qwen2.5-VL-32B](/models/qwen2-5-vl-32b) | Alibaba | Open Weight | 32K | 39.7 | Estimated | 28.13–51.16 |
| 71 | [Qwen2.5-72B](/models/qwen2-5-72b) | Alibaba | Open Weight | 128K | 37.3 | Supported | 20.44–54.20 |
| 72 | [DeepSeek R1 Distill Qwen 32B](/models/deepseek-r1-distill-qwen-32b) | DeepSeek | Open Weight | 128K | 34.2 | Estimated | 25.62–42.74 |
| 73 | [Qwen2.5 Coder 32B Instruct](/models/qwen2-5-coder-32b-instruct) | Alibaba | Open Weight | 128K | 34 | Supported | 18.61–49.43 |
| 74 | [Kimi K2](/models/kimi-k2) | Moonshot AI | Proprietary | 128K | 26.1 | Supported | 15.73–36.53 |
| 75 | [MiniMax M1 80k](/models/minimax-m1-80k) | MiniMax | Proprietary | 80K | 24.2 | Supported | 14.03–34.34 |

## Key Takeaways

- Top model: [Kimi K3](/models/kimi-k3) with a score of 74.9 and Supported evidence
- Best open-weight option: [Qwen3.8 Max](/models/qwen3-8-max) at #2
- Models included: 75

## Compare the Leaders

- [Kimi K3 vs Qwen3.8 Max](/compare/kimi-k3-vs-qwen3-8-max)