AI BENCHY Category
Coding Ranking
See which AI models perform best on Coding, which ones stay reliable, and where the biggest gaps appear. Sort by: Response Time (avg) ↑.
| Rank | Model | Company | Coding Score | Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #73 | GPT-5 Mini medium | OpenAI | 10.0 | 6.9 | 2/2 | 30.7s |
| #17 | Grok 4.20 Beta medium | X AI | 10.0 | 8.2 | 1/1 | 31.4s |
| #29 | Hy3 preview medium | Tencent | 10.0 | 7.8 | 1/1 | 31.4s |
| #46 | Claude Sonnet 4.6 medium | Anthropic | 6.9 | 7.6 | 1/2 | 33.9s |
| #74 | Laguna M.1 medium | Poolside | 4.3 | 6.9 | 0/1 | 35.6s |
| #126 | Kimi K2.5 none | Moonshot AI | 6.8 | 5.3 | 1/2 | 36.0s |
| #118 | Nemotron 3 Nano Omni 30b A3b Reasoning medium | NVIDIA | 3.3 | 5.4 | 0/1 | 38.1s |
| #9 | Gemini 3.5 Flash none | 8.2 | 8.9 | 1/2 | 39.6s | |
| #106 | Owl Alpha none | Openrouter | 7.0 | 5.7 | 1/2 | 39.7s |
| #121 | Mistral Small 4 medium | Mistral | 5.1 | 5.4 | 0/2 | 44.8s |
| #111 | gpt-oss-120b medium | OpenAI | 3.9 | 5.6 | 0/2 | 47.2s |
| #94 | GPT-5 Nano medium | OpenAI | 5.4 | 6.1 | 0/2 | 47.8s |
| #80 | DeepSeek V4 Pro high | DeepSeek | 2.8 | 6.6 | 0/2 | 51.8s |
| #59 | Qwen3.6 Flash medium | Qwen | 5.1 | 7.4 | 0/2 | 51.9s |
| #28 | GLM 5 Turbo medium | Z.ai | 7.3 | 7.9 | 1/2 | 53.9s |