MiniMax M1 Models by MiniMax
Last refreshed 2026-10-19. Next refresh: weekly.
Details
Capabilities
About
MiniMax-M1 is MiniMax's open-weight reasoning model family, built on MiniMax-Text-01 with a hybrid Mixture-of-Experts design and lightning attention: about 456 billion total parameters, about 45.9 billion active per token, and a native context of 1 million tokens. MiniMax trained it with large-scale reinforcement learning, including a method it calls CISPO, and released versions with 40K and 80K thinking budgets. The lightning-attention design is meant to make long reasoning cheaper to run, and MiniMax positions the line for complex software engineering, tool use, and long-context work.
Decision facts
- Best fit
- reasoningJSON / Tool usestructured outputs
- Capability starting point
- MiniMax-M1-80k with 1M context
- Lowest tracked input
- MiniMax M1 · $0.400/1M · MiniMax
- Closest related family
- MiniMax Image
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 1M context, reasoning, and JSON / Tool use.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| MiniMax-M1-80k | Use when the workload needs 1M context. | 2025-06 | 1M context | Current |
| MiniMax-M1-40k | Use when the workload needs 1M context, reasoning, and JSON / Tool use. | 2025-06 | 1M contextreasoningJSON / Tool use | Current |
Release Timeline
2 release groupsReplaced By
Keep for legacy integrations; evaluate MiniMax M3 before new work.
Specifications(3 models)
| Model | Released | Context | Parameters | Reasoning | JSON / Tool use | Structured Outputs |
|---|---|---|---|---|---|---|
| MiniMax-M1-80k | 2025-06 | 1M | 456B total, 45.9B active (MoE) | No | No | No |
| MiniMax-M1-40k | 2025-06 | 1M | 456B total, 45.9B active (MoE) | Yes | Yes | No |
Available From(4 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| MiniMax-M1-80k | Novita AI | $0.55 | $2.2 | Serverless |
| MiniMax-M1-80k | Fireworks AI | $0.9 | $0.9 | Serverless |
