Information Retrieval on FRAMES
81.5RecallGPT-5.4
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-5.4Category=Frontier models2026.06 | 81.5 | 81.5 | 86.5 | |
| Opus-4.6Category=Frontier models2026.06 | 81.4 | 81.4 | 87.8 | |
| Sonnet-4.6Category=Frontier models2026.06 | 78.7 | 78.7 | 85.9 | |
| Harness-1Parameters=20B, Category=Open-source small models2026.06 | 71 | 71 | 74.9 | |
| Kimi-K2.5Category=Frontier models2026.06 | 69.7 | 69.7 | 84 | |
| Context-1Parameters=20B, Category=Open-source small models2026.06 | 63.8 | 63.8 | 78.7 | |
| Tongyi DeepResearchParameters=30B, Category=Open-source small models2026.06 | 56.1 | 56.1 | 65.4 | |
| GPT-OSSParameters=120B, Category=Frontier models2026.06 | 49.5 | 49.5 | 86.3 | |
| GPT-OSSParameters=20B, Category=Open-source small models2026.06 | 39.1 | 39.1 | 70.2 | |
| Search-R1Parameters=32B, Category=Open-source small models2026.06 | 32.9 | 32.9 | 32.9 | |
| Qwen3Parameters=32B, Category=Open-source small models2026.06 | 16.4 | 16.4 | 40.2 |