Nejumi 4 - jaster (2-shot) - JEMHopQA: leaderboard
JEMHopQA, Japanese explainable multi-hop question answering, from the llm-jp-eval jaster suite as run 2-shot by Nejumi Leaderboard 4 (Weights & Biases Japan). Score is character-level F1 x100 on Nejumi's sampled test split; higher is better.
Metric: Character F1 (x100). Source: nejumi.ai. Saturation forecast: Around August 2027. 145 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 71.93 | #71 |
| 2 | Gemini 3.1 Pro (Preview) | 71.8 | #64 |
| 3 | Gemini 3.5 Flash | 71.58 | #66 |
| 4 | GPT-5.6 Sol (Max) | 71.35 | #22 (GPT-5.6 Sol) |
| 5 | GPT-5.4 (2026-03-05) (xHigh) | 71.29 | #63 (GPT-5.4 (2026-03-05)) |
| 6 | GPT-5.4 (2026-03-05) (High) | 71.25 | #63 (GPT-5.4 (2026-03-05)) |
| 7 | Gemini 2.5 Pro | 69.22 | #159 |
| 8 | Gemini 3 Flash (Preview) | 69.05 | #90 |
| 9 | GPT-5.6 Terra (Max) | 68.89 | #40 (GPT-5.6 Terra) |
| 10 | Gemini 3.6 Flash | 68.67 | #57 |
| 11 | Claude Opus 4.5 (Thinking) | 67.69 | #87 (Claude Opus 4.5) |
| 12 | Muse Spark Open (xHigh) | 67.41 | #42 (Muse Spark Open) |
| 13 | Claude Opus 4.1 (Thinking) | 67.29 | #146 (Claude Opus 4.1) |
| 14 | Claude Opus 4.6 (Thinking) | 67.08 | #67 (Claude Opus 4.6) |
| 15 | Qwen 3.6 Max (Preview) (Thinking) | 66.11 |
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-jemhopqa · How It Works · Data refreshed daily, snapshot 2026-11-05.