Benchmark
AutomationBench
AutomationBench is a tool-use benchmark that evaluates AI agents on automating real-world workflows, testing their ability to orchestrate tools and…
- Modality
- text
- Categories
- reasoning, agents, tool_calling
- Openness
- unknown
- Source
- llm_stats
- Reported scores
- 12
Reported scores
llm_stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Claude Fable 5 | Anthropic | 0.174 | 2026-06-09 | evidence |
| Claude Opus 5 | Anthropic | 0.26 | 2026-07-24 | evidence |
| Claude Sonnet 5 | Anthropic | 0.135 | 2026-06-30 | evidence |
| DeepSeek-V4-Flash-0731 | DeepSeek | 0.251 | 2026-07-31 | evidence |
| DeepSeek-V4-Pro-0813 | DeepSeek | 0.318 | 2026-08-13 | evidence |
| Gemini 3.7 Flash | 0.304 | 2026-08-13 | evidence | |
| GLM-5.3 | Z.ai | 0.482 | 2026-08-14 | evidence |
| GPT-5.6 Luna | OpenAI | 0.149 | 2026-07-09 | evidence |
| GPT-5.6 Sol | OpenAI | 0.181 | 2026-07-09 | evidence |
| GPT-5.6 Terra | OpenAI | 0.152 | 2026-07-09 | evidence |
| Kimi K3 | Moonshot AI | 0.308 | 2026-07-16 | evidence |
| Qwen3.8 Max | Qwen | 0.273 | 2026-08-02 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.