MiniMax-M3

MiniMax-M3 is positioned within MiniMax's "LLM" category on the official site, listed alongside related text models such as M2.7 and M2.5. A dedicated product page exists at the MiniMax domain, and the release blog post carries the headline "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality — All in One Model," suggesting the model is being marketed toward ambitious coding and multimodal use cases rather than narrow specialty tasks.

The most concrete real-world signal for MiniMax-M3 comes from an NVIDIA developer forum thread titled "MiniMax M3 : NVFP4 for Quad DGX Spark," indicating that the community is actively exploring quantized deployments on multi-node GB10 hardware. That thread implies interest in efficient local or on-prem inference paths for MiniMax-M3, which is the kind of practical detail that matters to teams evaluating where the model can realistically be hosted beyond the flagship API surface.

ClinePasscline-pass/minimax-m3minimax

Open provider page Provider docs

Quick Info

Provider: ClinePass
Model key: cline-pass/minimax-m3
Release date: Jun 1, 2026
Last updated: Jun 1, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.30
Output token cost: $1.20

Limits

Output tokens: 512,000 tokens
Context window: 1,048,576 tokens

Transparent token rates

Compare MiniMax-M3 pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about MiniMax-M3

ClinePass

Coverage

MiniMax officially released MiniMax M3 on June 1, 2026, describing it as a frontier-level model for specialized coding and agentic tasks that uses MSA (MiniMax Sparse Attention), a new attention architecture the team proposed to enable long-context scaling. The launch blog states M3 supports context windows of up to 1M According to the same announcement, M3 shows significant coding improvements over M2, approaching the level of leading closed-source models in bugfix, frontend/backend development, and performance optimization, while its agentic capabilities target common office workflows such as search and Office-suite tasks and have

ClinePass

CoverageRelease Notes

The MiniMax platform release-notes index lists MiniMax-M3 under a June 1, 2026 entry as "the latest M-series language model for agentic reasoning, tool use, coding, multimodal chat input, and long-context tasks." This is a concise first-party confirmation that M3 is a discrete, released model in the M lineup, distinct Because the entry appears in the official MiniMax API documentation release feed, it serves as authoritative timestamp and capability-summary evidence rather than deep technical detail. Other models on the page (Music 3.0, H3, M2.7, etc.) should not be conflated with M3 — only the M3-specific entry is relevant to the s