Hi!
DIT Model Radar tracks new AI models daily: real models or confirmed codenames only, no tools, no hype.
MiniMax's own site hosts an M3.1-Flash-Preview page plus matching text-generation docs, covering coding, research and visual tasks inside MiniMax Code. The preview guide lays out migration notes from MiniMax-M3, including reasoning that is always enabled, with effort adjustable from Low to Xhigh but no way to turn it off. No public pricing or general-availability details appear yet.

What changed: MiniMax's agent tools site and platform docs now carry a preview model page for M3.1-Flash-Preview, moving past the earlier MiniMax-M3. The page presents it as a preview for coding, research and visual work, with a direct migration comparison against M3.
Status: Official MiniMax model page and docs, preview build with no launch post or pricing published
- MiniMax hosts an M3.1-Flash-Preview page on agent.minimax.io and lists the name in its platform text-generation docs
- The page describes MiniMax Code workflows for debugging software, financial analysis and turning visual references into interactive code
- The guide compares it with MiniMax-M3 and notes reasoning is always enabled and cannot be disabled
- Not confirmed: no public pricing, rate limits or general-availability date appear in the official material
Right now M3.1 stands as an officially documented preview rather than a priced, generally available release.
[ $1 LAUNCH SPECIAL ] MiniMax M3.1 on DIT 15× the tokens. Just $1. ≈39M input or ≈8.8M output tokens.* Unlocks at launch. *Illustrative from MiniMax-M2.7-Turbo: $0.38/M in + $1.7/M out. |

Top AI models. Lower costs.
Access GPT, Claude, Gemini, and more through one OpenAI-compatible API—and save up to 70% on AI API costs. Create your DIT account, get your API key, and start building.