AI Model Timeline

Tracking the accelerating release frequency of frontier AI models.

Mistral 3 Family

Mistral·

LLMsopen-weightFrontier

Overview

Flagship Mistral Large 3 (675B MoE) and efficient Ministral 3 edge models. Largest open MoE model from Mistral.

Capabilities and innovations

675B MoE flagship + edge variantsEnterprise and edge deploymentMulti-language supportAdvanced function callingLargest open MoE from MistralEdge-optimized Ministral variantsImproved MoE routing efficiency

Benchmarks

No benchmark values are recorded for this model.

Architecture and hardware

Parameters
675B total (MoE)
Estimated VRAM at Q4
~389 GB, Frontier class
Quantization formats
GGUF, GPTQ, AWQ
Recommended runtime
Ollama
License
Apache 2.0

Links

More from Mistral

Data curated by AI Model Timeline. See the methodology for admission criteria, benchmark eras and source priority.