MiniMax-M3.1-Flash leads AutomationBench at 73% but trails on Terminal-Bench
ArtificialAnlys · x · 2026-10-19
Artificial Analysis detailed MiniMax-M3.1-Flash's agentic benchmark results, which split sharply:
- AutomationBench-AA: 73%, ahead of GLM-5.3-Flash (max, 60%) and DeepSeek V4.1 Flash 1017 (max, 67%)
- AA-Briefcase: 1474 Elo, between GLM (1453) and DeepSeek (1570)
- GDPval-AA: 1458 Elo, behind GLM (1642) and DeepSeek (1746)
- Terminal-Bench v4.0: 26%, below both rivals (33% and 34%)
Takeaway: strong and cost-efficient at automated task execution, but not yet competitive on terminal operations and real-world work tasks.
More from Models
- Users want open-format export from Claude Docs as charts only save as images — BlackHC · 2026-10-20
- EMNLP study: LLM safety interventions drop up to 4.5x when distress is wrapped in delusion — MaartenSap · 2026-10-20
- Cybersecurity-focused VLM OrcaCyber-Nano-1.0 trends on Hugging Face — orcarouter · 2026-10-20
- 0.8B local model with 2,000 labels beats Jev's frontier API on banking data — kalyan_kpl · 2026-10-20
- New TTS benchmark tests 26 models across 30,564 requests: fastest at 46ms — mattturck · 2026-10-20
- Agents handle unreleased language Jai flawlessly, even compile-time execution — lucasmeijer · 2026-10-20