MultiNRC
Native multilingual reasoning across languages. Scale AI.
- Published by
- Scale AI SEAL
- Category
- Knowledge
- Index weight
- 0.5
- Models
- 42
- Data as of
- 9 Sept 2026
Scale AI SEAL Leaderboards.
- 1Muse Spark 1.165.6%
- 2GPT-5 Pro65.2%
- 3Gemini 3.1 Pro64.7%
- 4GPT-5.4 Pro62.3%
- 5Muse Spark59.0%
- 6Gemini 3 Pro59.0%
- 7GPT-5.458.3%
- 8Claude Opus 4.657.1%
- 9GPT-552.1%
- 10GPT-5.149.0%
- 11o3-pro49.0%
- 12Claude Opus 4.548.6%
- 13Claude Opus 4.648.3%
- 14o345.5%
- 15Gemini 2.5 Pro45.1%
42 of 42
| # | |||||
|---|---|---|---|---|---|
| 1 |  Muse Spark 1.1Meta | 65.6% | Muse Spark 1.1 | 64.9 | – |
| 2 |  GPT-5 ProOpenAI | 65.2% | – | 60.7 | – |
| 3 |  Gemini 3.1 ProGoogle | 64.7% | – | 63.9 | – |
| 4 | GPT-5.4 ProOpenAI | 62.3% | – | 64.2 | – |
| 5 | Muse SparkMeta | 59.0% | – | 65.4 | – |
| 6 | Gemini 3 ProGoogle | 59.0% | – | 61.6 | – |
| 7 | GPT-5.4OpenAI | 58.3% | – | – | – |
| 8 |  Claude Opus 4.6Anthropic | 57.1% | – | 54.1 | – |
| 9 | GPT-5OpenAI | 52.1% | – | 59.5 | – |
| 10 | GPT-5.1OpenAI | 49.0% | – | 55.2 | – |
| 11 | o3-proOpenAI | 49.0% | – | 54.4 | – |
| 12 | Claude Opus 4.5Anthropic | 48.6% | – | 60.6 | – |
| 13 | Claude Opus 4.6Anthropic | 48.3% | – | 55.3 | – |
| 14 | o3OpenAI | 45.5% | – | 52.9 | – |
| 15 | Gemini 2.5 ProGoogle | 45.1% | – | 54.1 | – |
| 16 | o3OpenAI | 44.5% | – | 53.6 | – |
| 17 | GPT-5.2OpenAI | 42.2% | – | 61.5 | – |
| 18 | Claude Opus 4.5Anthropic | 41.2% | – | 58.0 | – |
| 19 | Claude Opus 4.1Anthropic | 38.4% | – | 58.1 | – |
| 20 | Claude Sonnet 4.5Anthropic | 35.8% | – | 52.8 | – |
| 21 |  Kimi K2.5Moonshot AIopen | 35.2% | kimi-k2.5 | 56.1 | – |
| 22 | Claude Opus 4Anthropic | 33.9% | – | 54.4 | – |
| 23 | Claude Opus 4.1Anthropic | 29.7% | – | 52.6 | – |
| 24 | Claude Opus 4Anthropic | 29.0% | – | 50.7 | – |
| 25 | Claude Sonnet 4.5Anthropic | 28.1% | – | 52.3 | – |
| 26 | Claude 3.7 SonnetAnthropic | 27.8% | – | 52.1 | – |
| 27 |  DeepSeek R1DeepSeekopen | 27.6% | 0528 | 48.9 | – |
| 28 |  Qwen3 235B A22BAlibabaopen | 27.1% | thinking, 2507 | 51.7 | – |
| 29 | Gemini 3.1 Flash LiteGoogle | 25.0% | – | 54.3 | – |
| 30 | GPT-5 miniOpenAI | 23.9% | – | 55.8 | – |
| 31 | o4-miniOpenAI | 22.2% | – | 48.0 | – |
| 32 | GPT-4.1OpenAI | 21.2% | GPT-4.1 | 47.3 | – |
| 33 | Kimi K2Moonshot AIopen | 18.5% | – | 48.8 | – |
| 34 | Claude Sonnet 4Anthropic | 18.4% | – | 49.5 | – |
| 35 | GPT 5.1 InstantOpenAI | 18.3% | – | – | – |
| 36 | Qwen3 235B A22BAlibabaopen | 17.6% | – | 48.7 | – |
| 37 |  GLM-4.5Zhipu AIopen | 17.4% | glm-4p5 | 50.2 | – |
| 38 | gpt-oss-120bOpenAIopen | 15.2% | – | 47.6 | – |
| 39 | GPT-4oOpenAI | 12.4% | GPT-4o | 42.3 | – |
| 40 | gpt-oss-20bOpenAIopen | 10.4% | – | 47.1 | – |
| 41 | GLM-4.5-AirZhipu AIopen | 10.4% | – | 46.0 | – |
| 42 | Llama 4 MaverickMetaopen | 8.4% | – | 42.6 | – |