Ebisu - JF-TE - Hit Rate@10: leaderboard
Evaluates Japanese financial implicit commitment recognition (JF-ICR) and terminology extraction (JF-TE). This is the paper v1 result table, whose stated generation limit is 1,024; current repository task defaults differ. JF-ICR uses accuracy, JF-TE uses maximal-term F1 and hit rates at 1/5/10, and the mean combines five metrics. Published fractions are converted to percentages. Explicit zero scores are retained; malformed predictions can receive zero under the public evaluator. Source-specified dated checkpoints are retained. This series reports Ebisu - JF-TE - Hit Rate@10 using Hit Rate@10 (%).
Metric: Hit Rate@10 (%). Source: arxiv.org. Saturation forecast: Around February 2027. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 70B Instruct | 51.11 | #541 |
| 2 | Qwen 2.5 32B Instruct | 47.86 | #509 |
| 3 | Claude Sonnet 4.5 | 47.52 | #148 |
| 4 | Llama-3.3-Swallow-70B-Instruct-v0.4 | 47.01 | #540 |
| 5 | Kimi K2 | 46.8 | #252 |
| 6 | Llama 4 Scout | 45.13 | #648 |
| 7 | Qwen 3 235B A22B 2507 Instruct FP8 | 43.01 | #234 |
| 8 | GPT-4o | 42.55 | #357 |
| 9 | DeepSeek V3.1 | 41.21 | #275 |
| 10 | Ministral-3-14B-Instruct-2512 | 37.04 | #625 |
| 11 | Qwen 3 32B | 17.63 | #449 |
| 12 | Qwen 3 14B | 10.97 | #542 |
| 13 | Qwen-14B | 10.76 | #1153 |
| 14 | Qwen 3 8B | 2.33 | #678 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ebisu-jf-te-hit-rate-10 · How It Works · Data refreshed daily, snapshot 2026-11-05.