Nejumi 4 - jaster (0-shot) - JSICK: leaderboard
JSICK, Japanese translation of the SICK entailment dataset, from the llm-jp-eval jaster suite as run 0-shot by Nejumi Leaderboard 4 (Weights & Biases Japan). Score is exact match (%) 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 | Gemma 4 E2B (IT) | 87 | #779 |
| 2 | GPT-5 Nano (2025-08-07) (High) | 87 | |
| 3 | Qwen 3 Next 80B A3B (Thinking) | 85 | #312 |
| 4 | DeepSeek V4 Flash (Max) | 85 | #155 (DeepSeek V4 Flash) |
| 5 | Grok 3 Mini | 85 | #289 |
| 6 | GPT-5.6 Terra (Max) | 85 | #40 (GPT-5.6 Terra) |
| 7 | llm-jp-4.1-33B (High) | 85 | #256 (llm-jp-4.1-33B) |
| 8 | Inkling (Max) | 85 | #144 (Inkling) |
| 9 | gemma-4-E4B-it | 84 | #577 |
| 10 | DeepSeek V4.1 Flash (1017) | 84 | #27 |
| 11 | Qwen3-Swallow-8B-RL-v0.2 (Thinking) | 84 | #602 (Qwen3-Swallow-8B-RL-v0.2) |
| 12 | O1 (2024-12-17) (High) | 84 | #157 (O1 (2024-12-17)) |
| 13 | Gemini 2.5 Flash | 83 | #255 |
| 14 | Grok 3 | 83 | #313 |
| 15 | DeepSeek V4.1 Flash | 83 | #37 |
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-jsick · How It Works · Data refreshed daily, snapshot 2026-11-05.