MERA: leaderboard
MERA (mera.a-ai.ru), the AI Alliance Russia evaluation suite for Russian-language LLMs: total score aggregated over its fixed instruction-task set, human baselines published upstream.
Metric: Total score. Source: mera.a-ai.ru. Saturation forecast: Around May 2027. 633 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 0.86 | #67 |
| 2 | GPT-5.4 | 0.82 | #83 |
| 3 | GLM-5.1 | 0.8 | #109 |
| 4 | Qwen 3 235B A22B 2507 (Thinking) | 0.8 | #263 |
| 5 | Qwen 3.6 35B A3B | 0.79 | #238 |
| 6 | Qwen 3.6 27B | 0.79 | #182 |
| 7 | Qwen 3.5 27B | 0.78 | #192 |
| 8 | Qwen 3.5 122B A10B | 0.77 | #184 |
| 9 | Claude Opus 4.5 | 0.76 | #87 |
| 10 | Qwen 3 30B A3B 2507 (Thinking) | 0.75 | #379 |
| 11 | Kimi K2.5 | 0.75 | #152 |
| 12 | Qwen 3 32B | 0.73 | #449 |
| 13 | Qwen 3.5 35B A3B | 0.73 | #268 |
| 14 | Gemma 4 31B (IT) | 0.71 | #178 |
| 15 | Qwen 3 Next 80B A3B (Thinking) | 0.71 | #312 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mera · How It Works · Data refreshed daily, snapshot 2026-11-05.