What hardware do you need to run MiniMax M3.1 Flash?

MiniMax M3.1 Flash at UD-Q4_K_XL is 249 GB of weights, with a context ceiling of 1M tokens. 23B active out of about 428B, natively multimodal and built for coding and long agent runs. It reads images and video as well as text, and the vision encoder is a separate 1.7 GB download this page does not price.

How good is MiniMax M3.1 Flash, really?

On the Artificial Analysis Intelligence Index v4.3 it scores 43 (very verbose: 180M output tokens over the index against a median of 100M), which puts it in the Sonnet-class band. In the same league as the labs' mainstream models on this index. See the whole table.

Head to head: vs GLM-5.3-Flash, the next model down the leaderboard.

What MiniMax M3.1 Flash costs either way

Renting the same model costs $0.09 per million input tokens and $0.5 per million output (OpenRouter, cheapest of thirteen active endpoints: Relace at FP4, checked 2026-10-25). Buying a machine only beats that if you use it hard enough, for long enough, that the hardware price divides down below the rental bill. What a million tokens costs each way puts the two prices side by side.

The machines that miss MiniMax M3.1 Flash, and what they run

None of the machines here with a published price that are still sold holds MiniMax M3.1 Flash at any window, and none of them misses by a window: its 249 GB of weights are larger than the usable memory in every one of them, before a single token of key-value cache. The Mac Studio M5 Ultra, 256GB comes nearest, 58 GB short at 4k, the shortest window on the list. Qwen3.8 27B is the strongest model six of the eight hold, and it scores 34 on the intelligence index against this model's 43.

Two machines here do hold MiniMax M3.1 Flash at 32k, and neither of them is an answer to what to buy, which is the question the rest of this page asks: the Mac Studio M3 Ultra, 512GB, no longer sold, at the $9,499 it launched at, and the Mac Studio M5 Ultra, 512GB, on sale with no published price. They are sold with 512 GB of memory and hand a model 384 GB of it, against the 254 GB MiniMax M3.1 Flash needs at 32k.

One machine per family, the largest configuration in each that still misses, nearest first. Short by is the weights plus the key-value cache at 4k measured against the memory that machine's GPU can address, which that machine's own page gives. The last column is the highest-scoring current model the machine holds at 32k.

Price the Mac Studio M5 Ultra, 256GB on GLM-5.3-Flash instead

The specifics

Parameters

428B, of which 23B are active per token

Quantisation

UD-Q4_K_XL

Weights on disk

249 GB

KV cache

4.5 GB at 32k context — Every one of the 60 layers caches the whole context: 2,048 B for the K and V of 4 KV heads at 128 wide, plus 256 B for the one index key a token leaves in the sparse-attention side cache. Sparse attention picks 16 blocks of 128 tokens to read, so it saves arithmetic and not memory, and the full 1M window costs about 145 GB. The four-bit build is an unsloth dynamic quant made from an already-quantised release: MiniMax publishes this model with its weights in MXFP8 and its expert weights in NVFP4, and no 16-bit version of it. How weights and cache add up.

Maximum context

1M tokens

Licence

MiniMax Community License

Sources

MiniMaxAI/MiniMax-M3.1-Flash, unsloth/MiniMax-M3.1-Flash-GGUF