MiniMax M3.1 Flash vs. MiniMax M3: One is cheaper, but not faster or better
MiniMax released M3.1 Flash on October 19, 2026, claiming 70% cost reduction for agentic tasks and 3x faster inference than M3. Pricing is $0.15 per million input tokens and $0.60 output, compared to M3's $0.30 and $1.20 rates. Independent testing via AutomationBench-AA scored Flash at 72.5% on multi-step automation tasks. The author's hands-on testing across three agent tasks found no clear improvement in accuracy versus M3 despite cost advantages.
Key points
- M3.1 Flash priced at 50% of M3 rates ($0.15/$0.60 per million tokens)
- Claims 70% cost savings and 3x faster inference
- Third-party AutomationBench-AA score: 72.5% task completion
- Author's testing showed no accuracy improvement over M3
Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. The New Stack did the reporting; read the original for the full story.
Read the full story at The New Stack
