MiniMax M3 (428B Multimodal)
Deploy MiniMax M3, a 428B natively multimodal MoE with a 1M-token context, on multi-GPU Clore.ai rigs
M3's interesting property is not its size, it is its attention. MiniMax Sparse Attention (MSA) cuts per-token compute at 1M context to roughly 1/20 of grouped-query attention, giving MiniMax a claimed 9× faster prefill and 15× faster decode than M2 at full context. For long-document and long-video work, that is the difference between a model you can afford to fill the window on and one you cannot.
With only ~23B parameters active per token, decode speed on a quantized build is closer to a mid-size model than to a 428B one.
Key Specs
Parameters
428B total, ~23B active (MoE)
Architecture
60 layers, GQA + MiniMax Sparse Attention, native vision encoder
Modality
Text, image, video in → text out
Context
1,000,000 tokens
License
minimax-community (custom, commercially restricted)
Release Date
1 June 2026
Weights
BF16 854GB · Q2_K_XL GGUF 143GB · Q3_K_M GGUF 195GB
Primary Tooling
vLLM, SGLang, llama.cpp, Transformers
Requirements
UD-IQ1_M
128GB
6× RTX 5090 (186GB) — comfortable
UD-Q2_K_XL
143GB
6× RTX 5090 or 8× RTX 4090 (192GB)
UD-Q3_K_M
195GB
8× RTX 5090 (248GB)
UD-Q4_K_S
248GB
10× RTX 5090 (310GB) or 4× RTX PRO 6000 (380GB)
At the last snapshot there were 72 servers with ≥167GB of VRAM (26 free) and 54 at ≥192GB (14 free) — enough to place a Q2 or Q3 build, but not so many that you should assume one will be waiting.
Multimodal input needs the vision tower loaded alongside the language model. Budget a few extra GB over the quant size if you plan to feed it images or video, and prefer rigs with plenty of system RAM for video decoding.
Deploy with llama.cpp
Deploy with vLLM
Pre-quantized checkpoints: MiniMaxAI/MiniMax-M3-MXFP8 (official), nvidia/MiniMax-M3-NVFP4 for Blackwell, cyankiwi/MiniMax-M3-AWQ-INT4 for AWQ stacks.
Clore.ai GPU Recommendations
6× RTX 5090
186GB
Q2_K_XL
~$1.50–2.60/hr
8× RTX 4090
192GB
Q2_K_XL
~$1.12–3.36/hr
8× RTX 5090
248GB
Q3_K_M
~$2.00–3.50/hr
4× RTX PRO 6000
380GB
Q4_K_S
~$5.00/hr
Use Cases
Long-video understanding — native video input plus the sparse-attention cost curve
Million-token document analysis — contracts, codebases, archives in a single pass
Agentic coding and "cowork" tasks — what MiniMax optimised M3 for
Cheap decode at frontier scale — 23B active parameters means tokens come out fast for the model's size
Multimodal RAG — one model for text, screenshots and video frames
Check the licence before commercial use. M3 ships under minimax-community, not an OSI licence. If you need unrestricted terms, GLM-5.2 is MIT and Qwen3.8-27B is Apache 2.0.
Troubleshooting
trust_remote_code errors
M3 ships custom modelling code — pass --trust-remote-code in vLLM
Sparse attention kernels missing
Update vLLM/SGLang; MSA support landed after the June release
Video input fails
Check the rig has enough system RAM and CPU for decoding; video is decoded host-side
OOM at 1M context
The KV cache is smaller than GQA but not free — start at 262K and work up
Slow prefill
Enable chunked prefill; MSA's advantage shows at long context, not on short prompts
Next Steps
Predecessor: MiniMax M2.7 — the April coding MoE
MIT alternative at similar scale: GLM-5.2
Single-card alternative: Qwen3.8-27B
Rig sizing: GPU Pricing & Availability
Links
Rent a GPU: RTX 5090 · RTX 4090 · Marketplace
Last updated
Was this helpful?