MiniMax has released its M3.1 Flash model, demonstrating strong performance across various benchmarks, particularly in agent capabilities and coding tasks. This new model achieves top-tier scores in areas like Agents' Last Exam (ALE) and BankerToolBench, rivaling or surpassing established models such as GPT-6.1 Sol and Claude Opus-5. Notably, M3.1 Flash offers significantly reduced costs, with input tokens priced at 1 yuan per million and output at 4 yuan per million, making it a highly cost-effective option for complex workflows and enterprise applications. The release positions MiniMax as a strong competitor in the AI landscape, especially alongside DeepSeek's recent Flash model, challenging the dominance of larger, more expensive models from Silicon Valley. AI
IMPACT Accelerates the trend of high-performance, low-cost AI models, potentially driving wider enterprise adoption of agent-based workflows.
RANK_REASON New model release from a frontier AI lab (MiniMax) with detailed performance benchmarks and cost analysis. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- Agents' Last Exam (ALE)
- AutomationBench-AA
- BankerToolBench
- Claude Fable-5
- Claude Opus-5
- DeepSeek
- DeepSWE v1.1
- GLM 5.3
- GPT 5.6 "Sol"
- GPT-6.1 Sol
- M3.1 Flash
- Minimax
- Space Bunny Explorers
- V4.1 Flash
- Zapier
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →