Chinese artificial intelligence company MiniMax has released its M3.1 Flash model. According to reporting by The Decoder, the new model achieves a score of 42 on the Artificial Analysis Intelligence Index, representing a 13-point increase over its predecessor while operating at half the cost. The M3.1 Flash also scores 73.7 percent on the DeepSWE v1.1 coding benchmark at a cost of under $0.30 per task. In both pricing and performance metrics, it matches the Deepseek V4.1 Flash 1017.
The Escalating Pricing Pressure in Chinese AI
The release highlights an ongoing trend among Chinese artificial intelligence developers to drive down the cost of advanced model deployment while simultaneously improving benchmark results. As competition intensifies among domestic firms, pricing models continue to drop aggressively. This race for affordability mirrors broader market dynamics where efficiency gains are immediately passed on to consumers in a bid for market share.
Unanswered Questions About Deployment and Limits
Despite the released benchmark figures and pricing details, several technical specifications and operational limits remain unknown. The announcement did not specify the hardware infrastructure required to run the model, nor did it disclose detailed context window sizes, rate limits, or broader API availability. Future observations will need to determine how the M3.1 Flash performs under heavy commercial workloads and whether these low price points remain sustainable over time.
As the pricing war among Chinese AI developers continues, market watchers will be tracking how competing firms respond to MiniMax and Deepseek.
Lead image: Brett Sayles / Pexels
