MiniMax-M3 API pricing & setup
As of 2026-10-17, the cheapest MiniMax-M3 channel is OpenRouter: $0.3 input / $1.2 output per 1M tokens (blended $0.525); official ¥2.1 / ¥8.4. Context 1M.
Context: 1M · Max output: 512K · Modalities: text, image, video · Listed: 2026-05-31
Channel comparison (per 1M tokens)
| Channel | Blended | Input | Output | Cached | Access & payment (this channel) | Notes | Source |
|---|---|---|---|---|---|---|---|
| Official (China) ¥2.1 / ¥8.4 | $0.549 | $0.314 | $1.25 | $0.063 | ✓CN direct✗Alipay?ID·unverified✗Intl card conditions source2026-10-17 | ≤512K 输入；官方标注「永久五折」，主价格为折后价（刊例价 ¥4.20 / ¥16.80 / 缓存读取 ¥0.84）；>512K 输入翻倍 models.dev ✓ | platform.minimaxi.com 2026-10-17 |
| OpenRouter $0.3 / $1.2 | $0.525 | $0.3 | $1.2 | $0.06 | ✓CN direct✓Alipay✗ID✗Intl card conditions source2026-10-17 | Original vendor endpoint: Minimax Quantized min $0.23 / $0.96 (fp4, CoreWeave) | openrouter.ai 2026-10-17 |
Access & payment conditions are verified per channel (see the column above); unverified never counts as met.
Free & deals
- MiniMax-M3 permanently 50% off list (≤512K input: ¥2.10/¥8.40) · official
How to use MiniMax-M3 in Claude Code / Codex / OpenCode
Provider-specific setup not yet verified. Generic setup: guides
Cheaper alternatives
DeepSeek V4.1 Flash ($0.523) · Qwen3.7 Plus ($0.523) · GPT-5.6 Luna ($0.45) ·
Compare
MiniMax-M3 vs Kimi K2.6 · DeepSeek V4.1 Flash vs Qwen3.8 Flash · DeepSeek V4.1 Flash vs GPT-6 Luna · Compare any two →
Rankings: Cheapest LLM APIs · LLM APIs usable from mainland China · All MiniMax models
Price history
- 2026-10-17 · Official (China) · MiniMax-M3 listed · ¥2.1 / ¥8.4
FAQ
How much does the MiniMax-M3 API cost?
Cheapest channel (OpenRouter): $0.3 input / $1.2 output per 1M tokens, verified 2026-10-17.
Can I use MiniMax-M3 from China?
Official (China): direct yes, Alipay no, ID required unverified, overseas card no; OpenRouter: direct yes, Alipay yes, ID required no, overseas card no
Is MiniMax-M3 free?
MiniMax-M3 permanently 50% off list (≤512K input: ¥2.10/¥8.40)
What is MiniMax-M3's context window?
1M tokens