MT-Bench PL - Coding: leaderboard
The "Coding" subtask of MT-Bench PL. The Polish-language version of MT-Bench, scoring multi-turn conversational and instruction-following ability via LLM-as-judge.
Metric: Judge Score (0-10). Source: huggingface.co. Saturation forecast: Estimated already saturated. 50 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Mistral Small 3.1 | 8.3 | #610 |
| 2 | Gemma 3 12B (IT) | 8.25 | #664 |
| 3 | Gemma 3 27B (IT) | 8.1 | #523 |
| 4 | Qwen 2.5 32B Instruct | 7.95 | #498 |
| 5 | Mistral Small 3 | 7.95 | #627 |
| 6 | Qwen 2 72B Instruct | 7.8 | #576 |
| 7 | Phi-4 | 7.6 | #706 |
| 8 | Gemma 2 27B (IT) | 7.45 | #645 |
| 9 | Llama 3.1 405B Instruct | 7.25 | #457 |
| 10 | Mistral-Small-Instruct-2409 | 7.1 | #702 |
| 11 | Mistral Large 2 (Jul) | 6.75 | #484 |
| 12 | Qwen 2.5 14B Instruct | 6.7 | #647 |
| 13 | Mixtral 8x22B | 6.45 | #766 |
| 14 | Llama 3.1 70B Instruct | 6.25 | #561 |
| 15 | Bielik-11B-v2.3-Instruct | 6.25 | #693 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mt-bench-pl-coding · How It Works · Data refreshed daily, snapshot 2026-10-20.