Business Utility Eval: leaderboard
June2026 Business Utility using Codex at the source’s declared xhigh effort over five attempts. Native0–1 mean task score is discounted by variability, rather than interpreted as pass rate; Claude Code and Gemini CLI have separate protocol boards.
Metric: Business Utility (%). Source: deepsense.ai. Saturation forecast: Around April 2028. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 42 | #37 |
| 2 | GPT-5.5 | 23 | #31 |
| 3 | GPT-5.4 | 22 | #79 |
| 4 | Claude Sonnet 4.6 | 13 | #90 |
| 5 | Gemini 3.1 Pro (Preview) | 12 | #58 |
| 6 | Gemini 3.5 Flash | 4 | #60 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=business-utility-eval · How It Works · Data refreshed daily, snapshot 2026-10-20.