MiniMax M2.5 Lightning for OpenClaw: Pricing, Setup, and What It's Good At

Current as of March 2026. M2.5 Lightning is the same 1M context, $0.30/$2.40 pricing story as M2.1 Lightning, built on the M2.5 architecture. The use case is narrow: you need to ingest a lot of data and get a short answer back. The 8K output ceiling is the limiting factor — keep that constraint in mind before you commit to it.

Specs

Provider	MiniMax
Input cost	$0.30 / M tokens
Output cost	$2.40 / M tokens
Context window	1M tokens
Max output	8K tokens
Parameters	N/A
Features	function_calling, reasoning

What it’s good at

1M Context, Cheap Input

$0.30/M for a million-token window is the main reason to use this model. Ingesting large codebases or full documentation sets without RAG costs under a dollar per run.

Function Calling

It follows the OpenAI tool-calling spec closely. I haven’t seen frequent argument hallucinations on schemas with five or fewer properties.

Where it falls short

8K Output Cap

A million tokens in, 8K tokens out. If your task needs a long response, this isn’t the right model. No exceptions.

Reasoning Drops on Complex Logic

It handles straightforward reasoning but drifts on deeply nested multi-step problems. Don’t use it for anything requiring precise logical chain-of-thought.

Best use cases with OpenClaw

• Large Repository Q&A — Load an entire codebase and ask targeted questions. Short answers fit the 8K cap.
• High-Volume Summarization — Process thousands of long transcripts daily. The pricing holds up at scale better than most alternatives.

Not ideal for

• Long Reports or Code Generation — The 8K output ceiling kills this use case. Look at M2.5 (non-lightning) instead.
• Real-time User Interfaces — TTFT climbs noticeably as you fill the 1M window.

Run it through Haimaker

Skip juggling API keys. One Haimaker key gives you access to every model on the platform. Tell OpenClaw:

Add Haimaker as a custom provider to my OpenClaw config. Use these details:

- Provider name: haimaker
- Base URL: https://api.haimaker.ai/v1
- API key: [PASTE YOUR HAIMAKER API KEY HERE]
- API type: openai-completions

Add the auto-router model:
- haimaker/auto (reasoning: false, context: 128000, max tokens: 32000)

Create an alias "auto" for easy switching. Apply the config when done.

Or skip model selection entirely — Haimaker’s auto-router picks the best model for each task so you don’t have to.

OpenClaw setup

Configure OpenClaw to use the Haimaker endpoint at api.haimaker.ai/v1 with your specific API key. Ensure the model ID is exactly minimax/MiniMax-M2.5-lightning to avoid routing to older M2 versions.

{
  "models": {
    "mode": "merge",
    "providers": {
      "minimax": {
        "baseUrl": "https://api.haimaker.ai/v1",
        "apiKey": "YOUR-MINIMAX-API-KEY",
        "api": "openai-completions",
        "models": [
          {
            "id": "MiniMax-M2.5-lightning",
            "name": "MiniMax M2.5 Lightning",
            "cost": {
              "input": 0.3,
              "output": 2.4
            },
            "contextWindow": 1000000,
            "maxTokens": 8192
          }
        ]
      }
    }
  }
}

How it compares

• vs Gemini 1.5 Flash — Gemini has a 1M context window too, but the pricing gets complicated at high output volumes. M2.5 Lightning is more predictable per token.
• vs GPT-4o-mini — 4o-mini has better reasoning but 128K context. If you need more context than that, it’s not an option.

Bottom line

The only reason to choose this over M2.5 is the 1M context window. If you don’t need that much input, the standard M2.5 is cheaper overall and removes the 8K output ceiling problem.

TRY MINIMAX M2.5 LIGHTNING ON HAIMAKER

---

For setup instructions, see our API key guide. For all available models, see the complete models guide.