SECQUE: leaderboard
SEC filing retrieval and question-answering benchmark for financial RAG over public company disclosures.
Metric: Strict Accuracy (imported). Source: huggingface.co. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 69 | #343 |
| 2 | Llama 3.3 70B Instruct | 65 | #534 |
| 3 | GPT-4o Mini | 64 | #599 |
| 4 | Qwen 2.5 32B Instruct | 61 | #498 |
| 5 | Phi-4 | 56 | #706 |
| 6 | Llama 3.1 8B Instruct | 48 | #1035 |
| 7 | Mistral Nemo Instruct (2407) | 46 | #936 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=secque · How It Works · Data refreshed daily, snapshot 2026-10-20.