AutomationBench (Public) leaderboard
As of November 14, 2026, among the results listed here, StepFun Step 5 Preview has the highest reported AutomationBench (Public) score, 44 (High), published by StepFun, ahead of LiquidAI LFM3-1.7B at 41.3. 6 models have a AutomationBench (Public) result in our records.
Scores as the model makers published them, not re-run by us; each row links to its announcement.
6 models ranked · 3 makers
| Rank | Model | Score | Setting | Reported by | Date |
|---|---|---|---|---|---|
| 1 | Step 5 Preview | 44 | High | StepFun | October 8, 2026 (model release) |
| 2 | LFM3-1.7B | 41.3 | Not published | Liquid | November 6, 2026 (model release) |
| 3 | DeepSeek V4 Flash Vision Exp | 25.7 | Not published | DeepSeek | August 21, 2026 (model release) |
| 4 | DeepSeek V4 Pro 0813 | 12.8 | Preview | DeepSeek | July 31, 2026 (model release) |
| 5 | DeepSeek V4 Flash 0731 | 10.8 | Preview | DeepSeek | July 31, 2026 (model release) |
| 6 | LFM2.5-2.6BRetired | 0.9 | Not published | Liquid | November 6, 2026 (model release) |
Other reported runs
| Model | Score | Setting | Reported by | Date |
|---|---|---|---|---|
| DeepSeek V4 Flash 0731 | 25.1 | Not published | DeepSeek | July 31, 2026 (model release) |
| DeepSeek V4 Pro 0813 | 31.8 | Not published | DeepSeek | August 12, 2026 (model release) |
How much to trust these numbers
- Source
- Reported by the model makers themselves. Not re-run by LLMBenchmark or anyone independent.
- Settings
- Settings differ between makers. Check the Setting column before comparing two scores.
- Spread
- 44 to 0.9 across 6 models.
- Freshness
- Newest result published November 6, 2026; oldest July 31, 2026.
What AutomationBench (Public) measures
AutomationBench (Public) appears in the release announcements of DeepSeek, Liquid, and StepFun. We have not found a public write-up of the benchmark itself yet.