Mistral Large (latest)

Mistral Large is positioned as a top-tier model from Mistral AI, a French company founded in April 2023 by former Meta Platforms and Google DeepMind researchers. The company set out to combine commercial AI products with a steady cadence of open-source large language models, beginning with the influential Mistral 7B released in September 2023. Within that lineage, Mistral Large is presented as Mistral's flagship offering aimed at advanced reasoning, multilingual understanding, and code generation, areas where it is meant to rival leading closed competitors such as GPT-4 and Claude 2.

The model's design emphasizes practical deployment for developers and data scientists, with documented support for on-demand API usage, function calling, and integration into agentic workflows. Third-party tutorials and guides describe Mistral Large as a strong fit for crafting code, handling multilingual tasks, and powering agent-based applications that require reliable instruction following. Because the supplied evidence consists of early 2024 tutorials that predate the current variant, detailed architecture, parameter counts, and benchmark figures for the November 2024 release are not confirmed here, and its strengths are best characterized by its stated emphasis on reasoning, code, and multilingual versatility within the broader Mistral family.

Mistralmistral-large-latestmistral-large

Open provider page Provider docs

Quick Info

Provider: Mistral
Model key: mistral-large-latest
Release date: Nov 1, 2024
Last updated: Dec 2, 2025
Knowledge cutoff: 2024-11
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.50
Output token cost: $1.50

Limits

Output tokens: 262,144 tokens
Context window: 262,144 tokens

Transparent token rates

Compare Mistral Large (latest) pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Mistral Large (latest)

Mistral

Official sourceAnnouncement

The most powerful AI platform for enterprises. Customize, fine-tune, and deploy AI assistants, autonomous agents, and multimodal AI with open models.

Mistral

Official sourceRelease Notes

Mistral's official changelog shows recent API and model updates into mid-2026, including the May 27, 2026 launch of Vibe — a unified agent at chat.mistral.ai with three modes (Work, Code, Chat) plus a CLI, VS Code extension, and Vibe Code Web in a managed cloud sandbox — and a documentation restructure that redirects / On the OCR side, the changelog confirms OCR 4.1 reached General Availability on August 30, 2026, that `mistral-ocr-latest` and `mistral-ocr-4` now point to `mistral-ocr-4-1`, and that the OCR API gained an `include_blocks` parameter returning paragraph-level bounding boxes with structural labels (text, title, list, tab

Mistral

CoverageBenchmark

This is Mistral AI's flagship model, Mistral Large 2 (version `mistral-large-2407`). $2 per million input tokens, $6 per million output tokens. 128,000 token context window. Includes independent benchmarks from Artificial Analysis.

Mistral

Coverage

... mistral-large-latest endpoint, and Mistral have updated that to point to the latest version of their Large model. Ollama now have mistral-large quantized to ...

Mistral

CoverageBenchmark

Compare Claude Opus 4.5 vs Mistral Large 3 (675B Instruct 2512): input $5/M vs $0.5/M, output $25/M vs $1.5/M tokens. Mistral Large 3 (675B Instruct 2512) is 1400% cheaper overall. Full API cost breakdown, context window, and benchmark comparison.