Nejumi 4 - jaster (2-shot) - MAWPS: leaderboard
MAWPS arithmetic word problems in Japanese, from the llm-jp-eval jaster suite as run 2-shot by Nejumi Leaderboard 4 (Weights & Biases Japan). Score is exact match of the final number (%) on Nejumi's sampled test split; higher is better.
Metric: Exact match (%). Source: nejumi.ai. Saturation forecast: Estimated already saturated. 145 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 99 | #255 |
| 2 | Gemini 3.5 Flash | 99 | #66 |
| 3 | Gemini 3.6 Flash | 99 | #57 |
| 4 | Claude Opus 5 (Max) | 99 | #16 (Claude Opus 5) |
| 5 | GPT-5.6 Luna (Max) | 99 | #79 (GPT-5.6 Luna) |
| 6 | GLM-5.2 (Max) | 99 | #80 (GLM-5.2) |
| 7 | DeepSeek-R1-Distill-Qwen-32B-Japanese | 99 | #665 |
| 8 | llm-jp-4.1-32B-a3B (High) | 99 | #366 (llm-jp-4.1-32B-a3B) |
| 9 | Gemini 3.1 Pro (Preview) | 98 | #64 |
| 10 | GPT-4.1 Mini | 98 | #363 |
| 11 | Gemini 3 Pro (Preview) | 98 | #71 |
| 12 | Grok 4 | 98 | #189 |
| 13 | Gemma 4 31B (IT) | 98 | #178 |
| 14 | Gemini 3 Flash (Preview) | 98 | #90 |
| 15 | Claude 3.5 Sonnet | 98 | #362 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=nejumi-4-jaster-2-shot-mawps · How It Works · Data refreshed daily, snapshot 2026-11-05.