models · XiaomiMiMo

MiMo-V2.5

Catalogued from llm-stats Open LLM Leaderboard at Quarantine — unverified, pending attestation.

What this registry has checked

Quarantine. We hold the publisher’s own published hash for these weights and have recorded it against this entry. We have not downloaded and hashed the weights ourselves, so the hash is the publisher’s claim, not our measurement.

Method	hf-published-sha256

Re-check this yourself → Source → Full record →

What is recorded

Publisher	XiaomiMiMo
Origin	Not recorded. The harvest wrote its own source name into this field, which tells us nothing about jurisdiction.
Licence	MIT
Parameters	310,775,040,000 (310.8B)
Context window	1,050,000 tokens
Modality	text+image+audio+video->text
Architecture	mimo_v2
Layers	48
Hidden size	4096
Attention	GQA (64 heads, 4 KV)
Native precision	F8_E4M3
Pulls recorded here	6
Downloads reported upstream	314,572

Reported scores

GPQA Diamond	84.9
IFBench	67.1
Humanity's Last Exam	27.2

Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.

Measured on real hardware

None of these runs are ours. Each figure is the reporter’s, linked to their own evidence, under their own trust label. The number shown is single-stream decode — one user, concurrency 1 — because aggregate throughput across many users is a different quantity and reads as far faster than anyone will see.

Reported to local-ai-registry

Hardware	Engine	Build measured	Decode, 1 stream	Label	Reported by
Radeon AI PRO R9700 ×3 devices	llama.cpp	IQ2_XXS	33.9 tok/s @ 32,768	candidate	localmaxxing
Measured outside 2K–32K context — decode speed depends heavily on context, so these are not comparable with the rows above
NVIDIA DGX Spark GB10	vllm	NVFP4	28.4 tok/s @ 500,000	candidate	localmaxxing
Ryzen AI Max+ 395	llama.cpp	IQ2_XXS	28.1 tok/s @ 768	candidate	localmaxxing
NVIDIA DGX Spark GB10	llama.cpp	Q4_K_M-GGUF	16.1 tok/s @ 1,000,000	candidate	localmaxxing

From local-ai-registry (MIT), commit 124959d: 18 runs across 5 machines, 0 validated — model revision and runtime pinned, launch accepted — and 18 candidate, their label for useful evidence without a reproducible-launch promise.

Elsewhere in the registry

All models → Browse with filters → What will run on my hardware? →

Registry record as of 2026-09-26. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.