Ambig-DS: leaderboard
As data-science agents shift from co-pilots to auto-pilots, silent misframing becomes a critical failure mode. Agents quietly commit to plausible but unintended task framings, producing clean, executable artifacts that hide their incorrect assessment of the task. Existing benchmarks score whether the pipeline runs, ignoring whether the agent recognized the task was underspecified. We introduce Amb
Metric: Full (self-reported). Source: benchmarklist.com. Saturation forecast: Around July 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 64 | #58 |
| 2 | Gemini 3 Flash | 63 | #99 |
| 3 | GPT-5.4 | 61 | #79 |
| 4 | Claude Haiku 4.5 | 56 | #280 |
| 5 | GPT-5.4 Nano | 49 | #328 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ambig-ds · How It Works · Data refreshed daily, snapshot 2026-10-20.