MiniMax launches M3.1 Flash: 3x faster, 1/3 the cost of M3, weights out Oct 23
MiniMax_AI · x · 2026-10-19
MiniMax officially announced M3.1 Flash, pitched as frontier-level intelligence at 1/3 the cost and 3x the speed of M3. The company says Space Bunny, which topped global leaderboards with 200T+ tokens, was an earlier version of this model.
Key benchmarks: Automation-Bench 62% and ALE 30% for Cowork, DeepSWE 74% for coding, and BabyVision 91% for visual reasoning, trading wins against DeepSeek-V4.1-Flash, Kimi-K3, GPT-5.6 Sol, and Opus 5.
API pricing is flat-rate: $0.15/$0.60 per 1M input/output tokens, with cache reads at $0.015/1M. Weights arrive Oct 23, alongside an M Plan subscription and the MiniMax Code desktop app.
More from Venture
- Sequoia backs Blacksmith to verify code as fast as agents write it — Sequoia Capital · 2026-10-19
- SK Hynix perps rally nearly 8% in 24 hours as AI chip stocks rebound — Polymarket · 2026-10-19
- HQ founder's postmortem: dropping hands-on onboarding for self-serve was a mistake—customers paid 10x for implementation — jacob_posel · 2026-10-19
- Lessons from buying from ~1,900 x402 APIs in one day: 1,255 delivered for $12.19 — SashSail · 2026-10-19
- Anthropic ARR re-accelerating as Opus 5.5 reverses Fable-era stall, but charts lag — robleclerc · 2026-10-19
- IIT Madras alumnus raises $9 million for AI semiconductor materials discovery startup — nordicinst · 2026-10-19