Skip to content

Mistral Large 3

Mistral Large 3, Mistral's open-weight 675B mixture-of-experts model with image input and a 256K-token context, joins the catalog.

Open Mistral Large 3
Grouped bar chart of base-model results for Mistral Large 3 against Deepseek-3.1 and Kimi-K2 on MMMLU, GPQA-Diamond, SimpleQA, AMC and LiveCodeBench

mistralai/mistral-large-3 is Mistral's open-weight multimodal mixture-of-experts model, the 2512 snapshot. It has 675B total and 41B active parameters, takes text and image input, supports tool calls and JSON output, and has a 256K-token context. It is released under Apache 2.0. Mistral Large 4mistralai/mistral-large-4 is the newer, larger model.

AI SDK
import { createAnyRouter } from "@anyr/ai-sdk-provider"
import { generateText } from "ai"

const anyrouter = createAnyRouter()
const { text } = await generateText({
  model: anyrouter("mistralai/mistral-large-3"),
  prompt: "Summarize the trade-offs of mixture-of-experts models in three bullets.",
})
console.log(text)
Coding agent
curl -fsSL https://anyr.dev/cli.sh | bash
anyr login
anyr claude --model "mistralai/mistral-large-3"

All routes are BYOK-only. Add a key for Token Harbor, OpenRouter, Nous Portal or Vercel under Dashboard → BYOK; the provider bills you at its own rates and AnyRouter charges nothing for a BYOK route.

Playbooks: AI SDK, OpenAI SDK, coding agents, playground.

Stacked win and lose bars for Mistral Large 3 Instruct against Deepseek V3.1 and Kimi K2 on general prompts and multilingual prompts, judged by humans via a third party
Mistral Large 3 Instruct preference comparison. Source: Mistral AI.

Looking for the full catalog? Browse all models.