whatmodel.sh / MiMo-V2-Flash (309B-A15B)

MiMo-V2-Flash (309B-A15B)

+ contribute

MiMo-V2-Flash (309B-A15B) is an open-weights language model from Xiaomi, released December 2025. Its published size is 309.8B parameters (15B active per token). Licence: mit.

Catalogue identity notes

Comparison-only, proposed so StepFun's own reproductions in the Step 3.5 Flash card count (column 'MiMo-V2 Flash', '15B' activated / '309B' total). Weights: https://huggingface.co/XiaomiMiMo/MiMo-V2-Flash (sha 1afd314a2406c282e0956375c34a676501c78649, created 2025-12-16, licence mit, 309,785,318,400 parameters in safetensors metadata); its card: 'MiMo-V2-Flash is a Mixture-of-Experts (MoE) language model with 309B total parameters and 15B active parameters.' Distinct from MiMo-V2-Flash-Base, MiMo-V2.5 and MiMo-V2-Pro.

66.2±1.2#127 of 719 LLMsfrom 12 benchmark results across 3 sources

$ how it compares

The LLMs ranked either side of MiMo-V2-Flash (309B-A15B), out of 828 with a score, then more from Xiaomi.

#	model	score	where	RAM	released
144	o4-mini OpenAI	66±1	api $1.93/M		2025-04
145	Qwen3.6 Flash Alibaba	66±1	api $0.56/M		2026-04
146	MiMo-V2-Flash (309B-A15B) Xiaomi	66±1	local	~108GB	2025-12
147	GLM-4.6 Z.ai	66±1	api $1.00/M +local	~58GB	2025-09
148	Kimi K2 Instruct (0711) Moonshot AI	66±1	api $1.00/M +local	~160GB	2025-07
	more from Xiaomi
15	MiMo-V2.6-Pro Xiaomi	79±1	api $0.54/M +local	~443GB	2026-09
27	MiMo-V2.6-Flash (309B-A15B) Xiaomi	77±1	api $0.17/M +local	~136GB	2026-09
91	MiMo-V2-Pro Xiaomi	70±1	comparison only		2026-03

see all LLMs · find the ones that fit your machine

$ what it's good at

agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #101 of 300 * · 19% of its score · 2 tests

Beats Grok 4.20 (0309 reasoning), Grok 4.3, o3-pro here, though they rank higher overall. Trails SmolLM3 3B, Nova 2 Pro (Preview), MiniMax-M2.1 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
BrowseComp-ZH	51.2% ✎	▁▁▁▁▁▃·▁▄▅▄█ beats 14/23	hard · avg 18%	▇▇▇▇ 10%	at its level
τ²-bench aggregate	84.1% ✎	▄······▄█·▄▄ beats 1/5	medium · avg 46%	▇▇▇▇ 9.5%	typical

coding▁▁▁▃▆██▇▆▄▃▁ #135 of 458 * · 31% of its score · 2 tests

Beats Qwen3.5 Plus, Gemini 2.5 Pro, Gemini 3.1 Flash-Lite here, though they rank higher overall. Trails Devstral Small 2 24B, Llama 3.1 Nemotron Ultra 253B v1, Nemotron 3.5 Lightning 30B-A3B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Arena WebDev (Code Arena)	1292.5 ✎	▁▂▄▃▅█▄▇▆▄▁▁ beats 19/119	avg 981	▇▇▇▇▇▇▇▇▇▇ 24%	sharp
LiveCodeBench v6 (window unspecified)	81.6% ✎	▂▃▄▄▅██▅▄▇▃▆ beats 46/65	medium · avg 48%	▇▇▇ 6.6%	small run

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #124 of 393 * · 14% of its score · 1 test

Beats LongCat 2.0 (1.6T-A48B) here, though they rank higher overall. Trails LFM2.5 2.6B, Nemotron-Cascade-8B, DeepSeek V3 0324 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
LMArena (text)	1386.3 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 226/379	avg 1268	▇▇▇▇▇ 14%	blunt

math▁▂▃▅▆▅▄█▄▁▁▁ #84 of 398 · 20% of its score · 4 tests

Beats Claude Sonnet 4.6, Claude Opus 4.5, MiniMax-M3 here, though they rank higher overall. Trails Granite 4.2 30B, INTELLECT-3, EXAONE 4.5 33B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
AIME 2025	95.1% ✎	▁·▁▁▁▂▅▆█▇▅▃ beats 129/133	easy · avg 67%	▇▇▇▇ 10%	sharp · maxed out · small run
IMO-AnswerBench	80.9% ✎	▁▁·▁▁▁▄▅█▄▂▃ beats 24/42	medium · avg 35%	▇▇ 5.3%	small run
HMMT November 2025	91.0% ✎	▁▁····▁▅▃█▇▁ beats 16/23	medium · avg 65%	▇ 2.5%	noisy
HMMT February 2025	95.4% ✎	▁▁▁▃▁▅▅▃██▁▄ beats 27/31	medium · avg 58%	▇ 1.9%	noisy · small run

web search██···█····██ #4 of 5 * · 10% of its score · 1 test

test	result	vs the others in its run	difficulty	counts	why
BrowseComp-ZH	51.2% ✎	▁▁▁▁▁▃·▁▄▅▄█ beats 14/23	hard · avg 18%	▇▇▇▇ 10%	at its level

other tests3 not in a task · 17% of its score

test	result	vs the others in its run	difficulty	counts	why
WildClawBench	30.8% ✎	▃·▅▅█·▅█▅▃▃▃ beats 3/17	very hard · avg 11%	▇▇▇ 6.8%	at its level
ResearchRubrics	54.3% ✎	███··██·█··█ beats 0/6	hard · avg 31%	▇▇▇ 6.4%	at its level
GAIA (no-file subset)	78.2% ✎	▃▃····█▃···▃ beats 5/6	medium · avg 47%	▇ 3.4%	at its level

$ where to run it

llama.cpp · recommended

version

Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
Windows
irm https://llama.app/install.ps1 | iex
other ways to install ↗
Run MiMo-V2-Flash (309B-A15B) · needs ~186.8 GB · from huggingface.co/unsloth
chat in the terminal
llama cli -hf unsloth/MiMo-V2-Flash-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf unsloth/MiMo-V2-Flash-GGUF:Q4_K_M

LM Studio

version

Install LM Studio · Open the app once before using the lms command.
download it (macOS / Windows / Linux) ↗
Run MiMo-V2-Flash (309B-A15B) · needs ~186.8 GB · from huggingface.co/unsloth
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/unsloth/MiMo-V2-Flash-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

MLX (Mac)

Install MLX (Mac) · Apple silicon Macs only, text only.
with uv
uv tool install mlx-lm
or
pip install mlx-lm
Run MiMo-V2-Flash (309B-A15B) · needs ~173.8 GB, Mac only · from huggingface.co/mlx-community
chat in the terminal
mlx_lm.chat --model mlx-community/MiMo-V2-Flash-4bit
Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model mlx-community/MiMo-V2-Flash-4bit

Weights and documentation: huggingface.co/XiaomiMiMo ↗

$ how the score was worked out · stepfun-ai/Step-3.5-Flash mod… 55%, LMArena leaderboard dataset 38%, WildClawBench: A Benchmark fo… 7%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
stepfun-ai/Step-3.5-Flash model card	9	▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 55%	Apache-2.0
LMArena leaderboard dataset	2	▇▇▇▇▇▇▇▇▇▇▇ 38%	CC BY 4.0
WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation (arXiv:2605.10912v1)	1	▇▇ 7%	CC BY 4.0
prior · pull toward the middle		▇ 1%
weighted average		= 66.2

$ improve this page

Update model information Add a place to run it Add a test result Propose a benchmark Correct a benchmark score Edit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated 2026-10-11 02:20 UTC · sources and licences · json