Introducing MiniMax M2.5 on Majestix AI
April 22, 2026by Majestix AImodelsminimaxcodingcodingagentic
Why MiniMax M2.5 matters
MiniMax M2.5 is a coding model in the Majestix lineup built for coding, agentic, tool use workloads, with 128K context and $0.32 / $1.00 per 1M tokens. This card is written for Majestix users, so the focus is on where the model fits inside our product rather than how to self-host it.
What it is good at
- Coding
- Agentic
- Tool Use
- Reasoning
Majestix platform snapshot
| Metric | Value |
|---|---|
| Provider | MiniMax |
| Category | Coding |
| Access | Guru+ |
| Context window | 128K |
| Max output | 32K |
| Pricing | $0.32 input / $1.00 output per 1M tokens |
Benchmarks and operating signals
| Signal | Value | Source |
|---|---|---|
| Context Window | 128K | platform limit |
| Max Output | 32K | platform limit |
| Output Price | $1.00 / 1M | Majestix pricing |
Where Majestix already uses it
- No public swarm templates currently pin this model. It is still available in the platform catalog.
When to choose it
MiniMax M2.5 is the right pick when you want coding and agentic without manually spelunking through vendor docs. On Majestix, you can compare it directly against the rest of the lineup on pricing, context, and swarm fit.