SQBench - L3 Business Scenarios: leaderboard
The business-scenario layer of SQBench (Shaqiu Community, 2026): the percent of its 60 L3 delivery tasks in finance, industry, healthcare and public administration that an agent completes with no triggered risk penalty, one run per task. Higher is better.
Metric: Strict Pass@1 (%; the 60 L3 domain business-scenario tasks on SQBench v1.0, 220 production-oriented delivery tasks in which an agent processes input assets, uses tools and produces a specified deliverable; a task passes only with functional Completion 1 and no triggered 10-dimension risk penalty; one run per configuration and task; higher is better). Source: arxiv.org. Saturation forecast: Around 2029. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.2 (Max) | 28.3 | #138 (GLM-5.2) |
| 2 | Kimi K3 | 26.7 | #54 |
| 3 | Claude Opus 4.8 (Max) | 25 | #56 (Claude Opus 4.8) |
| 4 | DeepSeek V4 Flash (High) | 23.3 | #258 (DeepSeek V4 Flash) |
| 5 | Kimi K2.6 | 21.7 | #178 |
| 6 | GLM-5.1 | 21.7 | #179 |
| 7 | Qwen 3.7 Plus | 21.7 | #207 |
| 8 | MiMo-V2.5 | 21.7 | #354 |
| 9 | DeepSeek V4 Pro (Max) | 21.7 | #188 (DeepSeek V4 Pro) |
| 10 | DeepSeek V4 Pro (High) | 21.7 | #188 (DeepSeek V4 Pro) |
| 11 | MiMo-V2.5-Pro | 20 | #279 |
| 12 | DeepSeek V4 Flash (Max) | 20 | #258 (DeepSeek V4 Flash) |
| 13 | Kimi K2.7 Code | 20 | #185 |
| 14 | MiniMax-M3 | 18.3 | #269 |
| 15 | MiniMax-M2.7 | 18.3 | #518 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=sqbench-l3-business-scenarios · How It Works · Data refreshed daily, snapshot 2026-11-13.