LLM Stats (NOVA-63): leaderboard
Amazon Nova multilingual evaluation covering 63 languages, used to compare broad language understanding and generation quality across those languages. Scores are reported by LLM Stats.
Metric: Score (%). Source: llm-stats.com. Source: qwen.ai. Saturation forecast: Around 2032. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 397B A17B | 59.1 | #146 |
| 2 | Claude Opus 4.6 (Thinking) | 59.1 | #63 (Claude Opus 4.6) |
| 3 | Qwen 3.7 Max | 59 | #75 |
| 4 | Qwen 3.7 Plus | 58.8 | #92 |
| 5 | Qwen 3.5 122B A10B | 58.6 | #173 |
| 6 | Qwen 3.5 27B | 58.1 | #187 |
| 7 | Qwen 3.6 Plus | 57.9 | #120 |
| 8 | Qwen 3.5 35B A3B | 57.1 | #256 |
| 9 | Kimi K2.6 (Thinking) | 56.7 | #104 (Kimi K2.6) |
| 10 | Qwen 3.5 9B | 55.9 | #379 |
| 11 | GLM-5.1 (Thinking) | 54.6 | #97 (GLM-5.1) |
| 12 | Qwen 3.5 4B | 54.3 | #475 |
| 13 | Qwen 3 Max (Thinking) | 54.2 | #208 (Qwen 3 Max) |
| 14 | DeepSeek V4 Pro (Reasoning, Max Effort) | 52.8 | #102 (DeepSeek V4 Pro) |
| 15 | Qwen 3.5 2B | 46.4 | #1148 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-nova-63 · How It Works · Data refreshed daily, snapshot 2026-10-20.