iScript-Bench: leaderboard
Metric: Function pass@1 (%) over all 116 iScript-Bench tasks: the generated Innovus Tcl script passes the sandbox syntax check (no ** ERROR in the tool log) and a Gemini-2.5-pro judge, given the requirement, a reference script and manual command definitions, confirms it fulfils the function; temperature 1; higher is better. Source: arxiv.org. Saturation forecast: Around May 2028. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 14.66 | #145 |
| 2 | Claude Sonnet 4.5 | 10.34 | #138 |
| 3 | GPT-4.1 | 2.59 | #240 |
| 4 | DeepSeek V3.1 | 2.59 | #260 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=iscript-bench · How It Works · Data refreshed daily, snapshot 2026-10-11.