AIDABench (Pass@1) - QA: leaderboard
Metric: Score (%) on the 226 question-answering tasks, a single run (pass@1), AIDABench's 603 end-to-end data-analytics tasks over real business files (mostly spreadsheets, up to 14 files per task); the model works in a Python sandbox through one execute_code tool (at most 20 calls) with an auxiliary spreadsheet summary, provider-recommended decoding; QA answers are judged pass/fail by QwQ-32B against the reference, charts get a 0-1 rubric score from a Gemini-3-Pro-Preview judge, spreadsheets pass/fail from a Claude-Sonnet-4.5 file evaluator; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 68.58 | #138 |
| 2 | Kimi K2 (Thinking) | 66.37 | #236 (Kimi K2) |
| 3 | Gemini 3 Pro (Preview) | 64.6 | #64 |
| 4 | DeepSeek V3.2 (Thinking) | 62.39 | #198 (DeepSeek V3.2) |
| 5 | Qwen 3 Max (2026-01-23) | 60.62 | #151 |
| 6 | DeepSeek V3.2 (Non-reasoning) | 60.18 | #198 (DeepSeek V3.2) |
| 7 | Qwen 3 235B A22B 2507 Instruct | 53.98 | #291 |
| 8 | Qwen 3 30B A3B 2507 (Thinking) | 52.65 | #366 (Qwen 3 30B A3B 2507) |
| 9 | Qwen 3 30B A3B 2507 Instruct | 35.4 | #464 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=aidabench-pass-1-qa · How It Works · Data refreshed daily, snapshot 2026-10-11.