llm-jp-3-13B-instruct3: benchmark results
LLM-jp 3 instruction-tuned release developed by NII's LLM-jp project, using Llama architecture and a 4,096-token context window. The named instruction-tuning generation remains distinct from the pretrained checkpoint and other instruct generations. Provider: National Institute of Informatics. Access: Open.
Aggregate Rating 1554 ± 41, rank #1655 of 10916 rated models, from 19 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| pfgen-bench - Completion Mode - Helpfulness | 0.47 | Helpfulness Score | 95.5 |
| pfgen-bench - Completion Mode - Score | 0.73 | pfgen Score (mean of three) | 93.6 |
| pfgen-bench - Completion Mode - Fluency | 0.82 | Fluency Score | 92.7 |
| pfgen-bench - Completion Mode - Truthfulness | 0.91 | Truthfulness Score | 89.6 |
| pfgen-bench - QA Mode - Fluency | 0.77 | Fluency Score | 83.6 |
| pfgen-bench - QA Mode - Helpfulness | 0.39 | Helpfulness Score | 80.6 |
| pfgen-bench - QA Mode - Score | 0.68 | pfgen Score (mean of three) | 80 |
| pfgen-bench - Chat Mode - Fluency | 0.67 | Fluency Score | 79.4 |
| pfgen-bench - QA Mode - Truthfulness | 0.88 | Truthfulness Score | 73 |
| pfgen-bench - Chat Mode - Helpfulness | 0.28 | Helpfulness Score | 72.1 |
| pfgen-bench - Chat Mode - Score | 0.59 | pfgen Score (mean of three) | 70.6 |
| YOMI-Bench - Kanji Reading QA (Multiple Readings) | 89.28 | Accuracy (%; yes/no judgement of whether a given reading is | 66.7 |
Interactive version: theaggregate.ai/model?slug=llm-jp-3-13b-instruct3 · How It Works · Data refreshed daily, snapshot 2026-11-08.