FITMYLLM · SEPTEMBER 6, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
230B
Parameters (10B active)
0
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 200K
Quant Bits VRAM @ 16K Quality IQ2_XXS 2.38 71.8 GB
68.9 + 2.9 KV
low IQ2_M 2.93 87.6 GB
84.7 + 2.9 KV
low Q2_K 3.16 94.2 GB
91.3 + 2.9 KV
low IQ3_XXS 3.25 96.8 GB
93.9 + 2.9 KV
low IQ3_XS 3.5 104.0 GB
101.1 + 2.9 KV
low Q3_K_S 3.64 108.0 GB
105.1 + 2.9 KV
low IQ3_M 3.76 111.5 GB
108.6 + 2.9 KV
low Q3_K_M 4 118.4 GB
115.5 + 2.9 KV
low Q3_K_L 4.3 127.0 GB
124.1 + 2.9 KV
moderate IQ4_XS 4.46 131.6 GB
128.7 + 2.9 KV
moderate Q4_K_S 4.67 137.7 GB
134.8 + 2.9 KV
moderate Q4_K_M 4.89 144.0 GB
141.1 + 2.9 KV
good Q5_K_S 5.57 163.5 GB
160.6 + 2.9 KV
good Q5_K_M 5.7 167.3 GB
164.4 + 2.9 KV
good Q6_K 6.56 192.0 GB
189.1 + 2.9 KV
excellent Q8_0 8.5 247.8 GB
244.9 + 2.9 KV
lossless FP16 16 463.4 GB
460.5 + 2.9 KV
lossless
Select your GPU above to see speed estimates and compatibility for each quantization.
Too big for a single GPU — plan a multi-GPU deployment
Even the lightest quant needs ~72 GB. Size GPUs, replicas, TCO and scaling for a production setup. Open in Enterprise →
▸ READY TO RUN THIS? RENT BY THE HOUR
RENT A GPU AND RUN MINIMAX-M2.7 NOW
Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.
Community Ratings Chat Coding Reasoning Creative Vision Roleplay Agentic
Loading ratings...
Run this model IQ2_XXS — 68.9 GB VRAM IQ2_M — 84.7 GB VRAM Q2_K — 91.3 GB VRAM IQ3_XXS — 93.9 GB VRAM IQ3_XS — 101.1 GB VRAM Q3_K_S — 105.1 GB VRAM IQ3_M — 108.6 GB VRAM Q3_K_M — 115.5 GB VRAM Q3_K_L — 124.1 GB VRAM IQ4_XS — 128.7 GB VRAM Q4_K_S — 134.8 GB VRAM Q4_K_M — 141.1 GB VRAM Q5_K_S — 160.6 GB VRAM Q5_K_M — 164.4 GB VRAM Q6_K — 189.1 GB VRAM Q8_0 — 244.9 GB VRAM FP16 — 460.5 GB VRAM
Ollama llama.cpp vLLM LM Studio KoboldCpp Jan Docker
▸ Easiest way to get started · Beginners
DOCS ↗ curl -fsSL https://ollama.com/install.sh | shCOPY
$ ollama run minimax:230b-q4_K_MCOPY
Tag may need adjustment — check ollama.com/library/minimax for available tags.
▸ SETUP GUIDE >_
Auto-setup with fitmyllm CLI Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.
Auto-detect GPU Live tok/s in chat Speed benchmarks 9 inference engines
GPUs that can run this model At Q4_K_M quantization. Sorted by minimum VRAM.
Find the best GPU for MiniMax-M2.7
Build Hardware for MiniMax-M2.7 ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
MiniMax-M2.7 — 230B MoE. ▸ SPECIFICATIONS
PARAMETERS 230B (10B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 200K tokens
CAPABILITIES chat, coding, reasoning, multilingual, math, agentic, tool_use
RELEASE DATE 2026-04-12
PROVIDER MiniMax
FAMILY minimax ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_XXS 2.38 68.9 GB 65% IQ2_M 2.93 84.7 GB 75% Q2_K 3.16 91.3 GB 78% IQ3_XXS 3.25 93.9 GB 82% IQ3_XS 3.5 101.1 GB 84% Q3_K_S 3.64 105.1 GB 85% IQ3_M 3.76 108.6 GB 86% Q3_K_M 4 115.5 GB 88% Q3_K_L 4.3 124.1 GB 90% IQ4_XS 4.46 128.7 GB 92% Q4_K_S 4.67 134.8 GB 93% Q4_K_M 4.89 141.1 GB 94% Q5_K_S 5.57 160.6 GB 96% Q5_K_M 5.7 164.4 GB 96% Q6_K 6.56 189.1 GB 97% Q8_0 8.5 244.9 GB 100% FP16 16 460.5 GB 100%
§ 01 BENCHMARK SCORES
GPQA Diamond 87.4
HLE 28.1
AA Intelligence 49.6
AA Coding 41.9
aa_ifbench 75.7
aa_terminal_bench 39.4
aa_tau2 84.8
aa_scicode 47.0
aa_lcr 68.7
§ 03 COMPATIBLE GPUs
14 @ Q4_K_M Feedback 