FREELINE free AI models, checked live
All models

MiniMax M3.1 Flash

Good
size 1M free on 1 of 6 routes Checked by listing

Free again on NVIDIA Build, one free route out of six on record.

Same as Strong coder class, a step under top tier

Made by MiniMax

About

MiniMax M3.1 Flash launched on 19 Oct 2026. Before that it ran without a name under the codename Space Bunny. It handles 1M tokens of context and is about three times faster than M3. Its weights were opened on 23 Oct 2026 at 428B total and 23B active.

Best for

  • Writing code
  • Long documents
  • Quick answers

Scores on record

Watch out

All routes

NVIDIA Build

hosted API Live

Free through NVIDIA Build's free endpoint. You need an NVIDIA account and phone verification.

Last checked 24 Oct 2026 at 18:30 IST

Use it free in 1 minute

  1. Open NVIDIA Build.
  2. Sign in, or ask us and we will walk you through it.
  3. Pick the model and send a test message.
  4. If it asks for a key, add yours in place of YOUR_KEY.

TokenRouter

free on account Paid now

Was free on a real user account on 20 Oct. The public price is $0.15 in, $0.60 out per 1M. Check your account.

Last checked 24 Oct 2026 at 18:30 IST

OpenCode Go

minimax-m3.1-flash-free Ended

Free on OpenCode Go until it dropped off the free list on 23 Oct.

Last checked 24 Oct 2026 at 18:30 IST

OpenRouter

minimax/minimax-m3.1-flash:free Ended

Free until 22 Oct. Now $0.15 per 1M input tokens and $0.60 per 1M output.

Last checked 24 Oct 2026 at 18:30 IST

Vercel AI Gateway

minimax/minimax-m3.1-flash Ended

Free until 22 Oct. Now a paid route.

Last checked 24 Oct 2026 at 18:30 IST

Command Code

free listing Ended

Was free. Now $0.15 in and $0.60 out per 1M.

Last checked 24 Oct 2026 at 18:30 IST

History

51 of 153 checks passed, 102 failed · listing checks, first one 21 Oct 2026 · last 20 shown

Each mark is one check we actually ran; the time is when it ran. A hollow mark is a failed check. No mark means no check in the log.