MiniMax has unveiled its latest AI model, the M3.1 Flash, which delivers significant improvements in both performance and cost-efficiency compared to its predecessor. According to The Decoder, the new model achieves a 13-point increase to reach a score of 42 on the Artificial Analysis Intelligence Index, while operating at half the previous cost.
The M3.1 Flash also matches the performance and pricing of Deepseek’s V4.1 Flash 1017, hitting 73.7 percent on the DeepSWE v1.1 coding benchmark for under $0.30 per task, as reported by The Decoder. This positions MiniMax competitively in the AI model market, balancing high accuracy with affordability.
For Japanese investors and technology firms, these advancements highlight the growing emphasis on cost-effective AI solutions that can support automation and data analysis across financial markets, including FX and equities trading.