English
chat.minimax_m2_5
chat.minimax_m2_5

MiniMax M2.5 is a further-tuned step in the M2 line for agentic workflows and coding — the same agent-and-coding DNA with tool loops and coding pushed a notch further than M2.1, at the same low price. Served on RouterBase over an OpenAI-compatible endpoint at $0.30 / 1M input and $1.20 / 1M output, with cached input at just $0.03, 10% under list; it sits between M2.1 and the latest M2.7.

Input
MiniMax M2.5
max_tokens4096
temperature1
top_p1
presence_penalty0
frequency_penalty0
MiniMax M2.5 Online
Hi! I'm a helpful AI assistant. What can I do for you?

MiniMax M2.5 API — The Further-Tuned M2 for Agents and Coding

The MiniMax M2.5 API serves MiniMax M2.5, a further-tuned step in the M2 line — stronger agentic tool loops and coding, still cheap.

MiniMax M2.5 continues the M2 line: the same agent-and-coding DNA, tuned another step. The MiniMax M2.5 API keeps the fast, cheap profile of the series — $0.30 per 1M input tokens and $1.20 per 1M output, with cached input at just $0.03 — while pushing tool-loop reliability and coding a notch further than M2.1. It sits between the earlier M2.1 and the latest M2.7: a capable middle for teams that want the newer tuning without chasing the very latest tag.

You reach the MiniMax M2.5 API through RouterBase on the OpenAI chat-completions protocol; set the model to minimax-m2-5 and your client is unchanged. The MiniMax M2.5 API bills $0.30 per 1M input, $1.20 per 1M output, and $0.03 for cached input — 10% under list, on one key that reaches the whole catalog.

Another step tuned

What the MiniMax M2.5 API gives you

The M2 agent-and-coding profile, pushed on.

Further-tuned M2

The MiniMax M2.5 API is a further-tuned step in the M2 line — same DNA, pushed on.

Agent tool loops

Reliable multi-step tool loops make the MiniMax M2.5 API a clean driver for agents.

Strong at coding

Writing and fixing code stays central for the MiniMax M2.5 API.

Low pricing

At $0.30 in and $1.20 out, the MiniMax M2.5 API is cheap enough for heavy agent loops — 10% below list.

Near-free cached reads

Cached input is just $0.03 per 1M, so repeated context on it costs almost nothing.

Drop-in across the line

Move between M2, M2.1, and the MiniMax M2.5 API by changing one string — the OpenAI shape is unchanged.

RouterBase dashboard preview
Get going

First call to the MiniMax M2.5 API in three steps

Quick to wire, ready to drive a tool loop.

  1. Create a key

    One RouterBase key reaches the MiniMax M2.5 API and every other model in the catalog.

  2. Point and name

    Send any OpenAI client to routerbase.com/v1 with model=minimax/minimax-m2.5. Streaming and function calling work out of the box.

  3. Wire your tools

    Give the MiniMax M2.5 API your function schemas and let it drive the loop; usage comes back per response so cost stays visible.

Get API Key Read docs
Pricing

Pay only for what you use

RouterBase passes through partner-tier pricing. Compared against the model's official published API rate.

Agents in production

Teams building on the MiniMax M2.5 API

The further-tuned agent-and-coding model, running cheap.

Selin KayaHead of Engineering, Bosphorus

Moving up to the MiniMax M2.5 API tightened our agent loops another notch without a rewrite.

Lars PetersenStaff Engineer, Kolding

The MiniMax M2.5 API handled our multi-step tasks a bit more cleanly than M2.1.

Ananya BoseCTO, Willowmark

Same $0.30 in, better tuning — the MiniMax M2.5 API was an easy move for us.

Kwesi BoatengFounder, Adinkra

Cached reads at $0.03 kept our long loops on the MiniMax M2.5 API basically free.

Elena PopovaBackend Lead, Nevsky

Function calling held up, so our whole tool layer runs on the MiniMax M2.5 API.

Ryo NakamuraML Lead, Tsuki Grid

Our coding agent got a little steadier again on the MiniMax M2.5 API.

Camille GirardPrincipal Engineer, Marelle

One string changed from M2.1; the MiniMax M2.5 API kept everything working.

Diego SalasFounder, Andino

We default agent and coding traffic to the MiniMax M2.5 API and keep it all on one key.

Hedda StrandEngineering Manager, Vik

The MiniMax M2.5 API was fast enough to sit inside our live coding assistant.

MiniMax M2.5 API — common questions

What to know before you route traffic to it.

RouterBase's hosted access to MiniMax M2.5, a further-tuned step in the M2 line for agentic workflows and coding, over an OpenAI-compatible endpoint.