Nejumi 4 - jaster (0-shot) - ALT English-to-Japanese: leaderboard
English-to-Japanese translation of Asian Language Treebank (ALT) parallel-corpus sentences, from the llm-jp-eval jaster suite as run 0-shot by Nejumi Leaderboard 4 (Weights & Biases Japan). Score is COMET (wmt22) x100 on Nejumi's sampled test split; higher is better.
Metric: COMET wmt22 (x100). Source: nejumi.ai. Saturation forecast: Estimated already saturated. 145 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 (Thinking) | 92.62 | #148 (Claude Sonnet 4.5) |
| 2 | Kimi K3 (Max) | 92.6 | #35 (Kimi K3) |
| 3 | Gemini 3.5 Flash | 92.58 | #66 |
| 4 | Claude Opus 4.1 (Thinking) | 92.58 | #146 (Claude Opus 4.1) |
| 5 | Claude Sonnet 4.6 (Thinking) | 92.56 | #94 (Claude Sonnet 4.6) |
| 6 | DeepSeek V4.1 Flash | 92.54 | #37 |
| 7 | Gemini 3.6 Flash | 92.54 | #57 |
| 8 | Claude Opus 4 (Non-reasoning) | 92.54 | #170 (Claude Opus 4) |
| 9 | Gemini 3.1 Pro (Preview) | 92.52 | #64 |
| 10 | DeepSeek V3 (0324) | 92.5 | #349 |
| 11 | Kimi K3 (Thinking) | 92.5 | #35 (Kimi K3) |
| 12 | Grok 4.5 (High) | 92.49 | #54 (Grok 4.5) |
| 13 | Claude Opus 4.5 (Non-reasoning) | 92.49 | #87 (Claude Opus 4.5) |
| 14 | GPT-4.1 | 92.48 | #258 |
| 15 | Claude Opus 4.6 (Thinking) | 92.48 | #67 (Claude Opus 4.6) |
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-0-shot-alt-english-to-japanese · How It Works · Data refreshed daily, snapshot 2026-11-05.