Mistral

Mistral Large 3

Name: Mistral Large 3
Author: Mistral

Frontier

HuggingFace

18.9KDownloads254LikesNov 2024Released256K tokensContextMistral Research LicenseLicense5 EntryQuality

Get started

— copy & paste to run locally

Quick specs

Parameters675B (41B active)

Architecturemoe (MoE)

Context256K tokens

Modalitytext+vision

Min RAM263.3 GB

Rec. RAM411.8 GB (Q4_K_M)

LicenseMistral Research License

FamilyMistral Large

✓ Code✓ Chat✓ Reasoning

About this model

Mistral-Large-Instruct-2411 is an advanced dense Large Language Model (LLM) of 123B parameters with state-of-the-art reasoning, knowledge and coding capabilities extending Mistral-Large-Instruct-2407 with better Long Context, Function Calling and System Prompt.

•Multi-lingual by design:: Dozens of languages supported, including English, French, German, Spanish, Italian, Chinese, Japanese, Korean,...
•Proficient in coding:: Trained on 80+ coding languages such as Python, Java, C, C++, Javacsript, and Bash. Also trained on more specific languages...
•Agent-centric:: Best-in-class agentic capabilities with native function calling and JSON outputting
•Advanced Reasoning:: State-of-the-art mathematical and reasoning capabilities
•Mistral Research License:: Allows usage and modification for non-commercial usages

Quantization options

VRAM estimates by quant level

No hardware detected — fit column shows raw VRAM estimates

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	263.3 GB	Low	—
Q3_K_S	3	330.8 GB	Low	—
NVFP4	4	378.0 GB	Medium	—
Q4_K_M	4	411.8 GB	Medium	—
Q5_K_M	5	486.0 GB	High	—
Q6_K	6	553.5 GB	High	—
Q8_0	8	722.3 GB	Very High	—
F16	16	1383.7 GB	Maximum	—

Hardware compatibility

Fit estimates across all hardware

Open calculator

Computing compatibility...

Memory breakdown

Reference: NVIDIA A10 24GB

Weights411.8 GB

KV Cache6.4 GB

Runtime2.4 GB

Headroom2.4 GB