FITMYLLM · AUGUST 17, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
675B
Parameters (39B active)
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 256K
Quant Bits VRAM @ 16K Quality IQ2_XXS 2.38 205.4 GB
201.3 + 4.1 KV
low IQ2_M 2.93 251.8 GB
247.7 + 4.1 KV
low Q2_K 3.16 271.2 GB
267.1 + 4.1 KV
low IQ3_XXS 3.25 278.8 GB
274.7 + 4.1 KV
low IQ3_XS 3.5 299.9 GB
295.8 + 4.1 KV
low Q3_K_S 3.64 311.7 GB
307.6 + 4.1 KV
low IQ3_M 3.76 321.9 GB
317.7 + 4.1 KV
low Q3_K_M 4 342.1 GB
338.0 + 4.1 KV
low Q3_K_L 4.3 367.4 GB
363.3 + 4.1 KV
moderate IQ4_XS 4.46 380.9 GB
376.8 + 4.1 KV
moderate Q4_K_S 4.67 398.6 GB
394.5 + 4.1 KV
moderate Q4_K_M 4.89 417.2 GB
413.1 + 4.1 KV
good Q5_K_S 5.57 474.6 GB
470.5 + 4.1 KV
good Q5_K_M 5.7 485.6 GB
481.4 + 4.1 KV
good Q6_K 6.56 558.1 GB
554.0 + 4.1 KV
excellent Q8_0 8.5 721.8 GB
717.7 + 4.1 KV
lossless FP16 16 1354.6 GB
1350.5 + 4.1 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 ~205 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 MISTRAL LARGE 3 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 — 201.3 GB VRAM IQ2_M — 247.7 GB VRAM Q2_K — 267.1 GB VRAM IQ3_XXS — 274.7 GB VRAM IQ3_XS — 295.8 GB VRAM Q3_K_S — 307.6 GB VRAM IQ3_M — 317.7 GB VRAM Q3_K_M — 338.0 GB VRAM Q3_K_L — 363.3 GB VRAM IQ4_XS — 376.8 GB VRAM Q4_K_S — 394.5 GB VRAM Q4_K_M — 413.1 GB VRAM Q5_K_S — 470.5 GB VRAM Q5_K_M — 481.4 GB VRAM Q6_K — 554.0 GB VRAM Q8_0 — 717.7 GB VRAM FP16 — 1350.5 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 mistral:675b-q4_K_MCOPY
Tag may need adjustment — check ollama.com/library/mistral 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
Find the best GPU for Mistral Large 3
Build Hardware for Mistral Large 3 Mistral Large 3 — 675B MoE flagship. Frontier-class with vision and tool use.
Read full model card ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
Mistral Large 3 — 675B MoE. ▸ SPECIFICATIONS
PARAMETERS 675B (39B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 256K tokens
CAPABILITIES chat, coding, reasoning, vision, tool_use
RELEASE DATE 2025-12-02
PROVIDER Mistral AI
FAMILY mistral ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_XXS 2.38 201.3 GB 65% IQ2_M 2.93 247.7 GB 75% Q2_K 3.16 267.1 GB 78% IQ3_XXS 3.25 274.7 GB 82% IQ3_XS 3.5 295.8 GB 84% Q3_K_S 3.64 307.6 GB 85% IQ3_M 3.76 317.7 GB 86% Q3_K_M 4 338.0 GB 88% Q3_K_L 4.3 363.3 GB 90% IQ4_XS 4.46 376.8 GB 92% Q4_K_S 4.67 394.5 GB 93% Q4_K_M 4.89 413.1 GB 94% Q5_K_S 5.57 470.5 GB 96% Q5_K_M 5.7 481.4 GB 96% Q6_K 6.56 554.0 GB 97% Q8_0 8.5 717.7 GB 100% FP16 16 1350.5 GB 100%
§ 01 BENCHMARK SCORES
MMLU-PRO 51.5
GPQA Diamond 68.0
LiveCodeBench 46.5
AIME 38.0
HLE 4.1
AA Intelligence 22.8
AA Coding 22.7
AA Math 38.0
aa_ifbench 36.2
aa_terminal_bench 15.9
aa_tau2 24.6
aa_scicode 36.2
aa_lcr 34.7
Feedback 