MERA - ruCodeEval: leaderboard
MERA is the Russian open benchmark run by the AI Alliance Russia; this is its ruCodeEval task: Russian-prompted HumanEval-style code generation, scored by unit tests. Each MERA task keeps its own native metric, so the boards are not on one shared scale.
Metric: pass@1 (%). Source: mera.a-ai.ru. Saturation forecast: Not forecast. 632 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.6 27B | 77.2 | #170 |
| 2 | Gemini 2.0 Flash | 76.89 | #342 |
| 3 | Qwen 3 235B A22B 2507 (Thinking) | 71.89 | #260 (Qwen 3 235B A22B 2507) |
| 4 | Gemma 4 31B (IT) | 71.71 | #176 |
| 5 | Qwen 3 4B 2507 (Thinking) | 71.71 | #538 (Qwen 3 4B 2507) |
| 6 | GPT-4o Mini | 70.43 | #599 |
| 7 | Qwen 3.6 35B A3B | 70.37 | #226 |
| 8 | Gemini 2.0 Flash Lite | 66.77 | #450 |
| 9 | Qwen 3 4B | 64.76 | #829 |
| 10 | Gemini 1.5 Flash | 63.05 | #618 |
| 11 | Gemini 1.5 Flash-8B | 61.28 | #819 |
| 12 | GigaChat 2 Pro | 59.88 | #682 |
| 13 | Qwen 3 14B | 58.35 | #537 |
| 14 | Qwen 3 32B | 57.68 | #443 |
| 15 | T-pro-it-2.0 | 54.27 | #380 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mera-rucodeeval · How It Works · Data refreshed daily, snapshot 2026-10-20.