MiniMax language model
MiniMax M3
MiniMax M3 is MiniMax's open-weight model (about 428B total, 23B active parameters) for coding and agent work, with image and video input. in this catalog it costs twice MiniMax M3.1 Flash, the newer October 2026 release.
01
price
input tokens
output tokens
cached input
per 1M tokens. octomind bills the listed provider price, with no markup, on every plan, including the free one (pay as you go from credits).
02
what real work costs
| workload | cost |
|---|---|
| one chat reply 3,000 tokens in, 500 out | $0.0015 |
| summarise a 100-page report 50,000 tokens in, 1,500 out, one request | $0.02 |
| 1,000 classification calls 500 tokens in, 20 out, per call | $0.17 |
| coding-agent session 2M tokens in across ~60 turns, 60,000 out, no cache | $0.67 |
| same session, cached 85% of input read from the prompt cache, the rest written to it | $0.26 |
at the listed rate. token counts are typical sizes, not measurements.
03
specs
- provider
- MiniMax
- context window
- 1M tokens — ≈ 1,500 pages of prose or 100,000 lines of code
- input
- text, images, video
- output
- text
- structured output
- no
- reasoning mode
- yes
- prompt caching
- yes — cache hits at $0.06/M
- weights
- open — MiniMax Community License
- released
- June 1, 2026
- status
- live
- model id
minimax-m3
integration notes
- temperature and top_p are ignored: M3 always uses temperature 1 and top_p 0.95.
- thinking is on by default; set thinking type to disabled or adaptive, and thinking tokens count toward max_tokens.
- in multi-turn calls, append the full assistant message back, preserving any <think> content in full.
04
benchmarks
as published by the maker; not measured by octomind.
05
when to pick it
pick it if
- you need image and video input at $0.30 / $1.20 per million
- you want open weights you can also self-host
- you reuse long prompts: cache hits bill at a fifth of input
skip it if
- MiniMax M3.1 Flash, at half the price, covers your workload
- you need JSON-schema constrained output
06
where it sits on price
blended price $0.525/M at 3 input tokens per output token: 14 live language models cost less, 29 cost more.
- Inkling 2 Small $0.475/M
- DeepSeek V4.1 Flash $0.525/M
- MiniMax M3 $0.525/M
- MiMo V2.6 Pro $0.544/M
- Inkling Small $0.675/M
07
use MiniMax M3 in octomind
one octohub login reaches every model in the catalog. pick this one per run, per session, or as your default.
octomind login
octomind run -m octohub:minimax-m3/model octohub:minimax-m3[model]
name = "octohub:minimax-m3"08
compare MiniMax M3
- MiniMax M3 vs Gemini 3.5 Flash Lite
- MiniMax M3.1 Flash vs MiniMax M3
- Beam vs MiniMax M3
- Inkling 2 Small vs MiniMax M3
09
similar models
nearest live models by listed input price.
- DeepSeek V4.1 Flash $0.3/M input
- Gemini 3.5 Flash Lite $0.3/M input
- Inkling 2 Small $0.3/M input
- GPT-OSS 120B $0.35/M input
10
MiniMax M3 faq
how much does MiniMax M3 cost?
MiniMax M3 lists at $0.3/M input, $1.2/M output, $0.06/M cached input per 1M tokens. a typical coding-agent session (2M tokens in, 60,000 out) comes to $0.67, or $0.26 when 85% of the input is read from the prompt cache. octomind bills the listed provider price, with no markup, on every plan, including the free one (pay as you go from credits).
what is the context window of MiniMax M3?
1M tokens, ≈ 1,500 pages of prose or 100,000 lines of code.
does MiniMax M3 accept images?
yes. it takes text, images and video as input and returns text.
is MiniMax M3 open source?
its weights are open under MiniMax Community License, so it can also be self-hosted.
when was MiniMax M3 released?
June 1, 2026, by MiniMax.
is MiniMax M3 free?
no. MiniMax M3 bills per token at $0.3/M input, $1.2/M output, $0.06/M cached input per 1M. octomind's free plan reaches it pay as you go from credits, so you pay only for the tokens you use, at the provider price. credits carry a 3% top-up fee.
how do i use MiniMax M3 in octomind?
run octomind login once, then octomind run -m octohub:minimax-m3. inside a running session, /model octohub:minimax-m3 switches to it without losing context.
use MiniMax M3 in octomind
choose a model when you start, then switch mid-session when the work changes.