MERA Code - CodeLinterEval: leaderboard
MERA Code (Russian) CodeLinterEval task: producing code a linter accepts, which penalises style-blind generation. Each task keeps its own metric — pass rates, CodeBLEU and chrF all appear across this suite — so read the metric label before comparing boards.
Metric: pass@1 (%). Source: mera.a-ai.ru. Saturation forecast: Estimated already saturated. 66 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 0.8B | 100 | #1424 |
| 2 | GPT-5.4 | 98.64 | #79 |
| 3 | MiniMax-M2.5 | 97.18 | #303 |
| 4 | GPT-5.1 Codex Max | 95.55 | #100 |
| 5 | GLM-5.1 | 93.45 | #97 |
| 6 | Qwen 3.6 27B | 93.45 | #170 |
| 7 | Claude Opus 4.6 | 93.09 | #63 |
| 8 | Qwen 3.6 35B A3B | 92.82 | #226 |
| 9 | Qwen 3.5 35B A3B | 89.64 | #256 |
| 10 | Qwen 3.5 9B | 88.18 | #379 |
| 11 | DeepSeek V4 Pro | 87.45 | #102 |
| 12 | Qwen 3.5 4B | 85.45 | #475 |
| 13 | MiniMax-M2.7 | 85.09 | #258 |
| 14 | Qwen 3.5 27B | 84.18 | #187 |
| 15 | Qwen 3.5 122B A10B | 80.36 | #173 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mera-code-codelintereval · How It Works · Data refreshed daily, snapshot 2026-10-20.