Information Retrieval on LongSealQA
89.1RecallHarness-1
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Harness-1Parameters=20B, Category=Open-source small models2026.06 | 89.1 | 89.1 | 94.4 | |
| Tongyi DeepResearchParameters=30B, Category=Open-source small models2026.06 | 85.5 | 85.5 | 85.5 | |
| GPT-5.4Category=Frontier models2026.06 | 82.4 | 82.4 | 88.5 | |
| Opus-4.6Category=Frontier models2026.06 | 82.3 | 82.3 | 87.6 | |
| Sonnet-4.6Category=Frontier models2026.06 | 78.3 | 78.3 | 84 | |
| Kimi-K2.5Category=Frontier models2026.06 | 66.1 | 66.1 | 91.1 | |
| Context-1Parameters=20B, Category=Open-source small models2026.06 | 56.3 | 56.3 | 85.8 | |
| Search-R1Parameters=32B, Category=Open-source small models2026.06 | 51.4 | 51.4 | 51.4 | |
| GPT-OSSParameters=120B, Category=Frontier models2026.06 | 46.4 | 46.4 | 91.4 | |
| Qwen3Parameters=32B, Category=Open-source small models2026.06 | 42.7 | 42.7 | 70.4 | |
| GPT-OSSParameters=20B, Category=Open-source small models2026.06 | 25.3 | 25.3 | 81.6 |