model family

MiniMax-M3

Published by MiniMaxAI. No machine on this network has measured this family yet. The members below are what the catalog knows exists; nothing about what they can do is claimed here until somebody runs one.

what it takes, what it returns
image + text → texttext → text

Declared by the network’s catalog, not measured — a signature says what this family would be asked. Whether it answered is the signed capability flags on each member.

0 of 41 builds measured · 41 still open
MiniMax-M3 BASE15 known builds · not measured
MiniMax-M3-MXFP4-AttnFP81 known build · not measured
MiniMax-M3-DSpark-GGUF6 known builds · not measured
MiniMax-M3-Alis-MLX-Dynamic1 known build · not measured
MiniMax-M3-DSpark1 known build · not measured
MiniMax-M3-Coder-Small1 known build · not measured
MiniMax-M3-MLX-mixed-3_6bit1 known build · not measured
MiniMax-M3-sm120-safe-GGUF3 known builds · not measured
MiniMax-M3-Mixed-4.5bit-MLX1 known build · not measured
Minimax-M3-v0-NVFP4-REAP251 known build · not measured
MiniMax-M3-ConfigI-MLX1 known build · not measured
MiniMax-M3-oQ41 known build · not measured
MiniMax-M3-EAGLE31 known build · not measured
MiniMax-M3-EAGLE3-GQA1 known build · not measured
MiniMax-M3-uncensored6 known builds · not measured

Builds this network knows of

Declared, not measured — this is what the catalog says exists and what it would cost a machine to hold. Nothing here is signed, and none of it claims the build works.

how much memory does your machine have

9 of 41 builds fit a 16 GB machine — budgeting 70% of it, or 11.2 GiB, because the cache and the runtime want the rest.

  • hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q2_KQ2_Kollamaderived by Anbeeld
    0.7 GiB
  • hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q3_K_MQ3_K_Mollamaderived by Anbeeld
    0.9 GiB
  • hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q4_K_MQ4_K_Mollamaderived by Anbeeld
    1.1 GiB
  • hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q5_K_MQ5_K_Mollamaderived by Anbeeld
    1.4 GiB
  • hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q8_0Q8_0ollamaderived by Anbeeld
    2.1 GiB
  • hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:BF16BF16ollamaderived by Anbeeld
    3.9 GiB
  • Inferact/MiniMax-M3-EAGLE3-GQABF16vllmderived by Inferact
    5.7 GiB
  • Inferact/MiniMax-M3-EAGLE3BF16vllmderived by Inferact
    6.1 GiB
  • nvidia/MiniMax-M3-DSparkFP32vllmderived by nvidia
    10.0 GiB
  • OsaurusAI/MiniMax-M3-Coder-SmallBF16mlxderived by OsaurusAI
    78.5 GiB
  • unsloth/MiniMax-M3-GGUF:IQ1_MIQ1_Mllamacppquantized by unsloth
    120.4 GiB
  • unsloth/MiniMax-M3-GGUF:IQ2_XXSIQ2_XXSllamacppquantized by unsloth
    125.6 GiB
  • spiritfather/MiniMax-M3-uncensored-GGUF:IQ2_SIQ2_Sllamacppderived by spiritfather
    134.0 GiB
  • spiritfather/MiniMax-M3-uncensored-GGUF:IQ3_SIQ3_Sllamacppderived by spiritfather
    148.0 GiB
  • unsloth/MiniMax-M3-GGUF:IQ3_XXSIQ3_XXSllamacppquantized by unsloth
    149.3 GiB
  • thetom-ai/MiniMax-M3-ConfigI-MLX4BITmlxderived by thetom-ai
    155.6 GiB
  • prometheusAIR/MiniMax-M3-sm120-safe-GGUF:IQ2_XSIQ2_XSllamacppderived by prometheusAIR
    159.9 GiB
  • pipenetwork/MiniMax-M3-MLX-3bit3BITmlxquantized by pipenetwork
    173.6 GiB
  • sparkarena/Minimax-M3-v0-NVFP4-REAP25BF16vllmderived by sparkarena
    174.2 GiB
  • pipenetwork/MiniMax-M3-MLX-mixed-3_6bit6BITmlxderived by pipenetwork
    178.0 GiB
  • avlp12/MiniMax-M3-Alis-MLX-Dynamic3BITmlxderived by avlp12
    179.4 GiB
  • prometheusAIR/MiniMax-M3-sm120-safe-GGUF:IQ3_XXSIQ3_XXSllamacppderived by prometheusAIR
    180.0 GiB
  • prometheusAIR/MiniMax-M3-sm120-safe-GGUF:Q3_KQ3_Kllamacppderived by prometheusAIR
    196.0 GiB
  • amd/MiniMax-M3-MXFP4-AttnFP8QUARKvllmderived by amd
    219.8 GiB
  • spectator2026/MiniMax-M3-AWQ-int4W4A16vllmquantized by spectator2026
    224.3 GiB
  • mlx-community/MiniMax-M3-4bit4BITmlxquantized by mlx-community
    224.9 GiB
  • unigilby/MiniMax-M3-oQ44BITmlxderived by unigilby
    228.1 GiB
  • unsloth/MiniMax-M3-GGUF:Q4_K_SQ4_K_Sllamacppquantized by unsloth
    231.5 GiB
  • nvidia/MiniMax-M3-NVFP4NVFP4vllmquantized by nvidia
    232.9 GiB
  • ressl/MiniMax-M3-uncensored-NVFP4NVFP4vllmderived by ressl
    242.4 GiB
  • spiritfather/MiniMax-M3-uncensored-GGUF:Q4_K_MQ4_K_Mllamacppderived by spiritfather
    246.1 GiB
  • unsloth/MiniMax-M3-GGUF:Q4_K_MQ4_K_Mllamacppquantized by unsloth
    246.7 GiB
  • Spangler3000/MiniMax-M3-Mixed-4.5bit-MLX4BITmlxderived by Spangler3000
    251.7 GiB
  • spiritfather/MiniMax-M3-uncensored-GGUF:Q5_K_MQ5_K_Mllamacppderived by spiritfather
    295.2 GiB
  • EmbeddedLLM/MiniMax-M3-FP8-dynamicFP8vllmquantized by EmbeddedLLM
    401.1 GiB
  • MiniMaxAI/MiniMax-M3-MXFP8MXFP8vllmpublished by MiniMaxAI
    413.3 GiB
  • pipenetwork/MiniMax-M3-MLX-8bit8BITmlxquantized by pipenetwork
    421.5 GiB
  • unsloth/MiniMax-M3-GGUF:Q8_0Q8_0llamacppquantized by unsloth
    422.5 GiB
  • spiritfather/MiniMax-M3-uncensored-GGUF:Q8_0Q8_0llamacppderived by spiritfather
    422.5 GiB
  • unsloth/MiniMax-M3-GGUF:BF16BF16llamacppquantized by unsloth
    794.3 GiB
  • MiniMaxAI/MiniMax-M3BF16vllmpublished by MiniMaxAI
    795.5 GiB

Sizes are what the catalog declares, not what a machine measured — fitting on disk is necessary and not sufficient. Nobody here has run these on your machine, and this narrows the candidates rather than certifying one.

the same builds, by member
MiniMax-M3 BASE
unsloth/MiniMax-M3-GGUF:IQ1_MIQ1_Mllamacpp120.4 GiB on diskquantized by unsloth
unsloth/MiniMax-M3-GGUF:IQ2_XXSIQ2_XXSllamacpp125.6 GiB on diskquantized by unsloth
unsloth/MiniMax-M3-GGUF:IQ3_XXSIQ3_XXSllamacpp149.3 GiB on diskquantized by unsloth
pipenetwork/MiniMax-M3-MLX-3bit3BITmlx173.6 GiB on diskquantized by pipenetwork
spectator2026/MiniMax-M3-AWQ-int4W4A16vllm224.3 GiB on diskquantized by spectator2026
mlx-community/MiniMax-M3-4bit4BITmlx224.9 GiB on diskquantized by mlx-community
unsloth/MiniMax-M3-GGUF:Q4_K_SQ4_K_Sllamacpp231.5 GiB on diskquantized by unsloth
nvidia/MiniMax-M3-NVFP4NVFP4vllm232.9 GiB on diskquantized by nvidia
unsloth/MiniMax-M3-GGUF:Q4_K_MQ4_K_Mllamacpp246.7 GiB on diskquantized by unsloth
EmbeddedLLM/MiniMax-M3-FP8-dynamicFP8vllm401.1 GiB on diskquantized by EmbeddedLLM
MiniMaxAI/MiniMax-M3-MXFP8MXFP8vllm413.3 GiB on diskpublished by MiniMaxAI
pipenetwork/MiniMax-M3-MLX-8bit8BITmlx421.5 GiB on diskquantized by pipenetwork
unsloth/MiniMax-M3-GGUF:Q8_0Q8_0llamacpp422.5 GiB on diskquantized by unsloth
unsloth/MiniMax-M3-GGUF:BF16BF16llamacpp794.3 GiB on diskquantized by unsloth
MiniMaxAI/MiniMax-M3BF16vllm795.5 GiB on diskpublished by MiniMaxAI
MiniMax-M3-MXFP4-AttnFP8
amd/MiniMax-M3-MXFP4-AttnFP8QUARKvllm219.8 GiB on diskderived by amd
MiniMax-M3-DSpark-GGUF
hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q2_KQ2_Kollama721 MiB on diskderived by Anbeeld
hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q3_K_MQ3_K_Mollama937 MiB on diskderived by Anbeeld
hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q4_K_MQ4_K_Mollama1.1 GiB on diskderived by Anbeeld
hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q5_K_MQ5_K_Mollama1.4 GiB on diskderived by Anbeeld
hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:Q8_0Q8_0ollama2.1 GiB on diskderived by Anbeeld
hf.co/Anbeeld/MiniMax-M3-DSpark-GGUF:BF16BF16ollama3.9 GiB on diskderived by Anbeeld
MiniMax-M3-Alis-MLX-Dynamic
avlp12/MiniMax-M3-Alis-MLX-Dynamic3BITmlx179.4 GiB on diskderived by avlp12
MiniMax-M3-DSpark
nvidia/MiniMax-M3-DSparkFP32vllm10.0 GiB on diskderived by nvidia
MiniMax-M3-Coder-Small
OsaurusAI/MiniMax-M3-Coder-SmallBF16mlx78.5 GiB on diskderived by OsaurusAI
MiniMax-M3-MLX-mixed-3_6bit
pipenetwork/MiniMax-M3-MLX-mixed-3_6bit6BITmlx178.0 GiB on diskderived by pipenetwork
MiniMax-M3-sm120-safe-GGUF
prometheusAIR/MiniMax-M3-sm120-safe-GGUF:IQ2_XSIQ2_XSllamacpp159.9 GiB on diskderived by prometheusAIR
prometheusAIR/MiniMax-M3-sm120-safe-GGUF:IQ3_XXSIQ3_XXSllamacpp180.0 GiB on diskderived by prometheusAIR
prometheusAIR/MiniMax-M3-sm120-safe-GGUF:Q3_KQ3_Kllamacpp196.0 GiB on diskderived by prometheusAIR
MiniMax-M3-Mixed-4.5bit-MLX
Spangler3000/MiniMax-M3-Mixed-4.5bit-MLX4BITmlx251.7 GiB on diskderived by Spangler3000
Minimax-M3-v0-NVFP4-REAP25
sparkarena/Minimax-M3-v0-NVFP4-REAP25BF16vllm174.2 GiB on diskderived by sparkarena
MiniMax-M3-ConfigI-MLX
thetom-ai/MiniMax-M3-ConfigI-MLX4BITmlx155.6 GiB on diskderived by thetom-ai
MiniMax-M3-oQ4
unigilby/MiniMax-M3-oQ44BITmlx228.1 GiB on diskderived by unigilby
MiniMax-M3-EAGLE3
Inferact/MiniMax-M3-EAGLE3BF16vllm6.1 GiB on diskderived by Inferact
MiniMax-M3-EAGLE3-GQA
Inferact/MiniMax-M3-EAGLE3-GQABF16vllm5.7 GiB on diskderived by Inferact
MiniMax-M3-uncensored
spiritfather/MiniMax-M3-uncensored-GGUF:IQ2_SIQ2_Sllamacpp134.0 GiB on diskderived by spiritfather
spiritfather/MiniMax-M3-uncensored-GGUF:IQ3_SIQ3_Sllamacpp148.0 GiB on diskderived by spiritfather
ressl/MiniMax-M3-uncensored-NVFP4NVFP4vllm242.4 GiB on diskderived by ressl
spiritfather/MiniMax-M3-uncensored-GGUF:Q4_K_MQ4_K_Mllamacpp246.1 GiB on diskderived by spiritfather
spiritfather/MiniMax-M3-uncensored-GGUF:Q5_K_MQ5_K_Mllamacpp295.2 GiB on diskderived by spiritfather
spiritfather/MiniMax-M3-uncensored-GGUF:Q8_0Q8_0llamacpp422.5 GiB on diskderived by spiritfather
ask your own gram

Your concierge can answer this from the network rather than from this page. The ask carries the operator id, so it resolves to a capability and an engine — not to a name somebody has to recognise.

What has this network measured about the MiniMax-M3 family (epn.family.minimax-m3), and can my gram run any of it? See /inferences/minimax-m3

open a gramx with this ask →epn.family.minimax-m3