MiniMax-M2 at a glance
MiniMax-M2 launched with open weights and an API priced at $0.30 input and $1.20 output per million tokens, which MiniMax put at 8% of Claude Sonnet's price at about twice the speed.
What changed with MiniMax-M2
On October 27, 2025, MiniMax open-sourced and launched MiniMax-M2, a model designed for agents and code. MiniMax said it handles long chains of tool calls across a shell, browser, Python interpreter and MCP tools, and that it ranked in the global top five on the Artificial Analysis intelligence benchmark at launch.
The API was priced at $0.30 per million input tokens and $1.20 per million output tokens, with an inference speed of about 100 tokens per second; MiniMax described this as 8% of Claude Sonnet's price at roughly twice the speed, and offered a free trial for a limited period. Weights were published on Hugging Face under a modified MIT license, and MiniMax upgraded its MiniMax Agent product to run on M2.
MiniMax-M2 benchmarks and pricing
Values are copied from the linked source on the date shown. They are not IA Compa measurements.
- Input price per 1M tokensas of · Source for Input price per 1M tokens: minimax.io (opens in a new tab)
- $0.30
- Output price per 1M tokensas of · Source for Output price per 1M tokens: minimax.io (opens in a new tab)
- $1.20
How it compares
Models that were current when this was published (released in the six months before it, newest version of each line), so these figures are from that time. Under each chart is the current best. (Models)
Current best (as of Nov 6, 2026): Claude Haiku 5.5, $0.10
Current best (as of Nov 6, 2026): Claude Haiku 5.5, $0.50
Where to compare MiniMax-M2
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
General leaderboards
Why MiniMax-M2 matters for businesses
For businesses, M2 showed that open models had become competitive for agent workloads at a fraction of frontier pricing, which lowers the cost of always-on automations, provided the license and data handling fit your requirements.
Sources for this story
- MiniMax M2 & Agent: Ingenious in Simplicity (opens in a new tab)Official
MiniMax · minimax.io
- MiniMaxAI/MiniMax-M2 (opens in a new tab)Official
Hugging Face · huggingface.co
- MiniMax-M2 model page (opens in a new tab)Benchmark
Artificial Analysis · artificialanalysis.ai
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #llm
- #minimax
- #open-weights
- #agents
- #coding
