MERA Code - stRuCom: leaderboard
MERA Code (Russian) stRuCom task: generating Russian-language code comments and docstrings. 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: chrF (%). Source: mera.a-ai.ru. Saturation forecast: Around 2033. 66 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 Codex Max | 38.3 | #100 |
| 2 | DeepSeek V4 Pro | 37.45 | #102 |
| 3 | GPT-5.4 | 34.59 | #79 |
| 4 | Qwen 3 Next 80B A3B (Thinking) | 33.36 | #320 (Qwen 3 Next 80B A3B) |
| 5 | Qwen 3 4B | 33.34 | #829 |
| 6 | Qwen 3 4B 2507 (Thinking) | 32.74 | #538 (Qwen 3 4B 2507) |
| 7 | Qwen 3 30B A3B 2507 (Thinking) | 32.69 | #376 (Qwen 3 30B A3B 2507) |
| 8 | Qwen 3 32B | 31.44 | #443 |
| 9 | Claude Opus 4.6 | 31.08 | #63 |
| 10 | Qwen 3.6 35B A3B | 30.59 | #226 |
| 11 | Qwen 3 8B | 30.51 | #674 |
| 12 | DeepSeek V4 Flash | 30.49 | #137 |
| 13 | Qwen 3.6 27B | 30.12 | #170 |
| 14 | GPT-4.1 | 29.66 | #248 |
| 15 | DeepSeek V3.1 | 29.31 | #264 |
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-strucom · How It Works · Data refreshed daily, snapshot 2026-10-20.