MiniMax

MiniMax M2.7

Type
language model
Context
205K tokens
Max output
131K
Released
18 March 2026
Knowledge cutoff
January 2025
API string
minimax/minimax-m2.7
Weights
open

Prices by provider

Provider Input, $ per 1M Output, $ per 1M In our data since
LLM Gateway $0.08 $0.32 31 Jul 2026
EmpirioLabs AI $0.15 $0.6 31 Jul 2026
AIHubMix coding-minimax-m2.7 $0.2 $0.2 31 Jul 2026
DeepInfra $0.25 $1 31 Jul 2026
OpenRouter $0.25 $1 31 Jul 2026
pinstripes $0.255 $0.55 31 Jul 2026
Pioneer $0.279 $1.2 31 Jul 2026
opencode-go $0.3 $1.2 31 Jul 2026
frogbot $0.3 $1.2 31 Jul 2026
auriko $0.3 $1.2 31 Jul 2026
vultr $0.3 $1.2 31 Jul 2026
alibaba-cn $0.3 $1.2 31 Jul 2026
novita-ai $0.3 $1.2 31 Jul 2026
minimax-cn $0.3 $1.2 31 Jul 2026
302.AI $0.3 $1.2 31 Jul 2026
AIHubMix minimax-m2.7 $0.3 $1.2 31 Jul 2026
Abacus.AI $0.3 $1.2 31 Jul 2026
FastRouter $0.3 $1.2 31 Jul 2026
Fireworks AI $0.3 $1.2 1 Aug 2026
Hugging Face $0.3 $1.2 31 Jul 2026
Kilo Code $0.3 $1.2 31 Jul 2026
Merge Gateway $0.3 $1.2 31 Jul 2026
MiniMax $0.3 $1.2 31 Jul 2026
NanoGPT $0.3 $1.2 31 Jul 2026
OpenCode Zen $0.3 $1.2 31 Jul 2026
OrcaRouter $0.3 $1.2 31 Jul 2026
Together AI $0.3 $1.2 31 Jul 2026
Vercel $0.3 $1.2 31 Jul 2026
ZenMux $0.306 $1.222 31 Jul 2026
CrossModel $0.33 $1.32 31 Jul 2026
hyper $0.42 $1.62 1 Aug 2026
Cortecs $0.47 $1.4 31 Jul 2026
SambaNova $0.6 $2.4 31 Jul 2026
unorouter $0.819 $3.276 31 Jul 2026

Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.

A single provider appears several times if it sells this model under different identifiers and at different prices — for example, at different weight precision. The identifier is shown next to the name. Identical offers that arrived from two sources under different spellings are merged into one row.

The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.

Measurement results

Task set Result Run conditions Measured by
Arena Score, programming 1,451.3 1,445–1,457
Arena Score, expert questions 1,437.81 1,428–1,447
Arena Score in English 1,427.74 1,422–1,433
Arena Score, hard prompts 1,422.37 1,418–1,427
Arena Score in French 1,420.92 1,405–1,437
Arena Score, long queries 1,420.78 1,415–1,426
Arena Score, mathematics 1,420.2 1,408–1,432
Arena Score, multi-turn dialogue 1,412.35 1,405–1,420
Arena Score in Spanish 1,409.44 1,393–1,426
Arena Score, instruction following 1,405.99 1,400–1,412
Arena Score, overall 1,405.32 1,401–1,409
Arena Score, web development 1,398.49 1,392–1,405
Arena Score in Russian 1,379.45 1,371–1,388
Arena Score, creative writing 1,353.29 1,345–1,361
Terminal-Bench — working in the command line 45.1 % · with a tuned harness https://www.tbench.ai/leaderboard/terminal-bench/2.0
WeirdML — unusual machine learning tasks 36.95 % https://htihle.github.io/weirdml.html
CritPt — physics problems 0.57 %
Arena Score, agent tasks -0.12 0–0