Information Retrieval on SEC filings
58.9RecallHarness-1
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Harness-1Parameters=20B, Category=Open-source small models2026.06 | 58.9 | 65.9 | 67.7 | |
| Opus-4.6Category=Frontier models2026.06 | 58.9 | 69.6 | 60.1 | |
| Context-1Parameters=20B, Category=Open-source small models2026.06 | 46.7 | 49.8 | 50.7 | |
| Kimi-K2.5Category=Frontier models2026.06 | 45.4 | 47.8 | 54.9 | |
| GPT-5.4Category=Frontier models2026.06 | 45.2 | 46.5 | 46.4 | |
| Tongyi DeepResearchParameters=30B, Category=Open-source small models2026.06 | 41.9 | 43.3 | 47.8 | |
| Sonnet-4.6Category=Frontier models2026.06 | 41.1 | 48 | 42.1 | |
| GPT-OSSParameters=120B, Category=Frontier models2026.06 | 39.5 | 50.4 | 48.8 | |
| GPT-OSSParameters=20B, Category=Open-source small models2026.06 | 13.3 | 16.6 | 25.2 | |
| Search-R1Parameters=32B, Category=Open-source small models2026.06 | 10.9 | 15.5 | 10.9 | |
| Qwen3Parameters=32B, Category=Open-source small models2026.06 | 5.8 | 7.2 | 21.5 |