MiniMax
8 models built · 8 served
MiniMax is a Shanghai-based lab, founded 2021, building the MiniMax-M model series along with audio and video generation tools. On OpenKey they offer 8 models, all paid text-generation models (no free tier from this lab), ranging from the original MiniMax-01 up through the newest release, MiniMax M3 (May 2026). The naming pattern (M1, M2, M2.1 through M2.7, M3) shows a lab iterating fast on a single core architecture rather than spinning up separate model families — each release tends to bring incremental context or pricing changes rather than a full redesign. Context windows top out at just over 1M tokens on M3 and the original MiniMax-01.
Lineup
For most text work, MiniMax M2.7 ($0.18/M input, $0.72/M output, 204,800 context) is the cheapest of the recent M2-series models and ranked in design_arena's dataviz benchmark (rank 19, elo 1271). If you need multimodal input (text, image, video) or the largest context window (1,048,576 tokens), go with MiniMax M3 at $0.30/M input and $1.20/M output — it's also the newest model and the top design_arena performer in this lineup (rank 11, elo 1295). MiniMax M2.5 is the cheapest overall at $0.12/M input and $0.48/M output if cost matters more than benchmark rank. Avoid MiniMax M1 unless you specifically need its 1,000,000-token context — it's priced highest in the lineup at $0.40/M input and $2.20/M output.
- MiniMax M3May 31, 2026 1.0M $0.309 $1.24 ctx 1.0M · in $0.309 · out $1.24
- MiniMax M2.7Mar 18, 2026 205K $0.185 $0.742 ctx 205K · in $0.185 · out $0.742
- MiniMax M2.5Feb 12, 2026 205K $0.124 $0.494 ctx 205K · in $0.124 · out $0.494
- MiniMax M2-herJan 23, 2026 66K $0.309 $1.24 ctx 66K · in $0.309 · out $1.24
- MiniMax M2.1Dec 23, 2025 205K $0.309 $1.24 ctx 205K · in $0.309 · out $1.24
- MiniMax M2Oct 23, 2025 205K $0.263 $1.05 ctx 205K · in $0.263 · out $1.05
- MiniMax M1Jun 17, 2025 1M $0.412 $2.27 ctx 1M · in $0.412 · out $2.27
- MiniMax-01Jan 15, 2025 1.0M $0.206 $1.13 ctx 1.0M · in $0.206 · out $1.13
Prices per 1M tokens, flat 3% fee included.
Models MiniMax serves
Beyond its own models, MiniMax hosts 8 models in the OpenKey catalog. Prices are MiniMax’s own list price per 1M input tokens, before our fee.
| Model | Built by | Context | Input /M |
|---|---|---|---|
| MiniMax-01 | MiniMax | 1.0M | $0.200 |
| MiniMax M1 | MiniMax | 1M | $0.400 |
| MiniMax M2 | MiniMax | 205K | $0.255 |
| MiniMax M2-her | MiniMax | 66K | $0.300 |
| MiniMax M2.1 | MiniMax | 205K | $0.300 |
| MiniMax M2.5 | MiniMax | 205K | $0.300 |
| MiniMax M2.7 | MiniMax | 205K | $0.300 |
| MiniMax M3 | MiniMax | 1.0M | $0.300 |
From the OpenRouter catalog, Jul 28, 2026. All providers
Privacy policy ↗Terms of service ↗
Questions
- What's the price range for MiniMax models on OpenKey?
- Input pricing runs from $0.12/M tokens (MiniMax M2.5) up to $0.40/M tokens (MiniMax M1). Output pricing runs from $0.48/M (M2.5) to $2.20/M (M1). OpenKey pricing is provider price × 1.03 — for example, M2.5's $0.12 input becomes $0.1236/M tokens with the flat 3% fee applied.
- Which MiniMax model should I use for the best price-to-performance ratio?
- MiniMax M2.7 is a good default: $0.18/M input, $0.72/M output, a 204,800-token context window, and rank 19 (elo 1271) on the design_arena dataviz benchmark. If you need lower cost and don't need benchmark-ranked output, MiniMax M2.5 is cheaper at $0.12/M input and $0.48/M output.
- Does MiniMax offer any free models on OpenKey?
- No. All 8 MiniMax models on OpenKey are paid — there's no free tier from this lab, unlike some of the other 25 free models available across OpenKey's full 329-model roster.