MiniMax M2.5
MiniMax M2.5
Version: 1
Fireworks•Last updated April 2026
MiniMax M2.5 is a Mixture-of-Experts model built for state-of-the-art coding, agentic tool use, and search, trained with reinforcement learning across hundreds of thousands of real-world environments.
Coding
Agents

Fireworks on Foundry

Models available for use with Fireworks on Foundry deliver optimized, best-in-class performance on the Fireworks Inference Cloud. Fireworks on Foundry is a Preview subject to Azure Preview terms and the following supplemental terms: When you use Fireworks on Foundry, data is shared between Microsoft and Fireworks AI, and different compliance and data handling rules will apply. See documentation for details. Customers are responsible for evaluating whether data sharing between Microsoft and Fireworks is appropriate for their organization’s compliance requirements.

Key capabilities

About this model

MiniMax M2.5 is a Mixture of Experts (MoE) language model built for state-of-the-art coding, agentic tool use, search, and office work. It was extensively trained with reinforcement learning across hundreds of thousands of real-world environments, enabling it to plan like an architect and generalize across unfamiliar scaffolding and tools. The model delivers significantly faster task completion, improved token efficiency, and exceptional cost-effectiveness, making it well-suited for production-scale agentic applications and complex, multi-step workflows.

Key model capabilities

  • State-of-the-art coding across 10+ languages (Go, C, C++, TypeScript, Rust, Kotlin, Python, Java, JavaScript, PHP, and more)
  • Agentic tool use with strong generalization across unfamiliar scaffolding
  • Deep search and information retrieval
  • Office work including Word, PowerPoint, and Excel financial modeling
  • Parallel tool calling for faster task completion
  • Efficient reasoning with optimized token usage

Use cases

See Responsible AI for additional considerations for responsible use.

Key use cases

  • Full-stack software development across the entire development lifecycle
  • Agentic workflows with tool calling and search
  • Document generation and office productivity
  • Expert-level research and information retrieval
  • Multi-step complex task automation

Out of scope use cases

The provider has not supplied this information.

Pricing

Pricing is based on a number of factors, including deployment type and tokens used. See pricing details here.

Technical specs

The provider has not supplied this information.

Training cut-off date

The provider has not supplied this information.

Training time

The provider has not supplied this information.

Input formats

Text

Output formats

Text

Supported languages

English

Sample JSON response

The provider has not supplied this information.

Model architecture

MiniMax M2.5 is a Mixture of Experts (MoE) language model developed by MiniMax. It was trained using the CISPO reinforcement learning algorithm with an agent-native RL framework called Forge.
PropertyValue
ArchitectureMixture of Experts (MoE)
Number of Experts256
Selected Experts per Token8
Number of Layers (Dense layer included)62
Number of Attention Heads48
Context Length196,608
Vocabulary Size200,064

Long context

Context Length: 192k tokens (196,608)

Optimizing model performance

The provider has not supplied this information.

Additional assets

The provider has not supplied this information.

Training disclosure

Training, testing and validation

The provider has not supplied this information.

Distribution

Distribution channels

The provider has not supplied this information.

More information

The provider has not supplied this information.
Model Specifications
Context Length196608
LicenseOther
Last UpdatedApril 2026
Input TypeText
Output TypeText
ProviderFireworks
Languages1 Language