MERA Code - YABLoCo: leaderboard
MERA Code (Russian) YABLoCo task: repository-scale long-context code completion. 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: Around 2035. 66 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 14.42 | #248 |
| 2 | Qwen 2.5 72B Instruct | 14.42 | #454 |
| 3 | Claude Opus 4.6 | 13.46 | #63 |
| 4 | Yi-Coder-9B-Chat | 13.46 | #1262 |
| 5 | Gemini 2.5 Flash | 12.02 | #245 |
| 6 | Qwen 2.5 Coder 32B Instruct | 11.06 | #528 |
| 7 | GPT-5.4 | 10.58 | #79 |
| 8 | DeepSeek V3.1 | 10.58 | #264 |
| 9 | Mixtral 8x22B Instruct (v0.1) | 10.58 | #756 |
| 10 | GPT-OSS-120B | 9.62 | #341 |
| 11 | GigaChat 2 Max | 9.13 | #536 |
| 12 | DeepSeek V4 Pro | 8.65 | #102 |
| 13 | GLM-5.1 | 8.17 | #97 |
| 14 | Qwen 3 Coder 480B A35B Instruct | 8.17 | #315 |
| 15 | Kimi K2 0905 | 7.69 | #288 |
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-yabloco · How It Works · Data refreshed daily, snapshot 2026-10-20.