DataSpace: leaderboard
Metric: Task accuracy (%; share of the 410 DataSpace tasks, each a question over a task-local heterogeneous workspace of CSV, JSON, SQLite, Markdown, PDF and video artifacts, whose returned table matches the gold table under the deterministic evaluator: header-invariant column alignment, type- and precision-aware normalization and order-aware row comparison; DataSpace-Agent reference harness unless the row names another harness). Source: arxiv.org. Saturation forecast: Around December 2026. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Grok 4.5 | 66.34 |
| 2 | GPT-5.6 Sol | 64.63 |
| 3 | Kimi K3 | 53.41 |
| 4 | MiMo-V2.5 | 39.27 |
| 5 | Claude Sonnet 5 | 32.93 |
| 6 | MiniMax-M3 | 28.54 |
Interactive version: theaggregate.ai/benchmark?slug=dataspace · How It Works · Data refreshed daily, snapshot 2026-09-29.