Head-on confrontation with DeepSeek! MiniMax M3.1 Flash released, output speed pushed to 200+ TPS, price cut to 1/2

What happened: MiniMax released M3.1 Flash, a lightweight model optimized for Coding and Agent tasks with a 1 million token context, two days after DeepSeek launched V4.1 Flash-1017. Key numbers: Input 1 yuan per million tokens, output 4 yuan per million tokens, cache read 0.1 yuan per million tokens. Why it matters: The two Flash models launched within 48 hours with close benchmark results, which the content describes as the first "dual strong"格局 for domestic models in the Flash segment. Pricing is about half of DeepSeek's peak rates and level with its idle rates. Coverage: 3 outlets in 24h

MiniMax released M3.1 Flash, a lightweight model optimized for Coding and Agent tasks with a 1 million token context, two days after DeepSeek launched V4.1 Flash-1017.

The two Flash models launched within 48 hours with close benchmark results, which the content describes as the first "dual strong"格局 for domestic models in the Flash segment. Pricing is about half of DeepSeek's peak rates and level with its idle rates.