Anthropic's developer API for the Claude model family, billed per million tokens with separate input, output and prompt-caching rates.
Usage-based pricing
Mistral AI API
Mistral AI's developer API (Studio) for its language, OCR, speech, embedding and moderation models, billed per token, per page or per minute depending on the service, with batch and cached-input discounts.
Overview
The Mistral API, managed through the Studio developer console, serves Mistral's generalist text and multimodal models (Mistral Large, Medium, Small and the smaller Ministral family) alongside OCR, speech transcription and text-to-speech, code completion, embedding, moderation and time-series models. The docs also list third-party open-weight models (Z.ai GLM) and an agents and workflows layer for building tool-using pipelines.
Text models are priced per million tokens, OCR per 1,000 pages and speech models per minute. The pricing page states a 50% reduction for batch processing and up to 90% lower input cost for cached tokens. Studio can be activated in a free mode to generate a first API key, but the pages fetched do not state free-tier limits or rate limits.
Only Mistral Large had a plainly stated token rate on the pricing page fetched; other model rates were not readable as text, so they are not recorded here. Several models are published as open weights (Apache 2.0 is named for some), and the vendor also sells consumer Vibe/Le Chat subscriptions that are outside this record.
Pricing
Usage-based pricing
Usage-based rates
| Item | Rate | Details | Source and check date |
|---|---|---|---|
| Mistral Large, input | USD 0.50 per 1M input tokens | Batch processing is 50% cheaper; cached input tokens up to 90% cheaperCurrency inferred from the US pricing page (shown as $ only). Page states 'Mistral Large costs $0.5 /M tokens in'. | Source for Mistral Large, input (opens in a new tab)Checked October 24, 2026 |
| Mistral Large, output | USD 1.50 per 1M output tokens | Batch processing is 50% cheaperCurrency inferred from the US pricing page (shown as $ only). Page states '$1.5 /M tokens out'. | Source for Mistral Large, output (opens in a new tab)Checked October 24, 2026 |
Prices checked on October 24, 2026 from the vendor's official page. Vendors change prices; confirm before buying. How we check prices. Official pricing page (opens in a new tab)
Who it's for
Developers and teams building applications on hosted language, document, speech and embedding models, including those who want open-weight models they can also run elsewhere.
Key features
Generalist text and multimodal models (Large, Medium, Small, Ministral) OCR with paragraph-level bounding boxes and confidence scores Speech transcription (including realtime and diarization) and text-to-speech Code completion and code embeddings (Codestral) Text embeddings, moderation and safety classifiers Agents, tools and durable workflows Batch processing and cached-input discounts Studio console with keys, playground, evaluations and SDKs
Pros and limitations
Limitations
Only the Mistral Large rate was readable as text on the pricing page; other model rates are not recorded Free-mode limits and rate limits are not stated on the pages fetched OCR and speech are billed per page and per minute rather than per token
Integrations
- Official SDK clients
- Mistral Compute (vendor-hosted infrastructure)
- Admin console with SSO and workspaces
Use cases
Chat and agentic applications with tool use Document OCR and retrieval pipelines Transcription and voice features Code completion and embeddings for search
Tags
How this page was researched
This page is based on desk research of the vendor's public pages and documentation. SIDAN Stack has not tested this product hands-on. How we research. Last verified: October 24, 2026.
Sources
- Mistral pricing (opens in a new tab) - Pricing page, checked October 24, 2026
- Mistral AI documentation (opens in a new tab) - Documentation, checked October 24, 2026
- Mistral models overview (opens in a new tab) - Documentation, checked October 24, 2026
Alternatives to Mistral AI API
Cloud platform for web scraping and browser automation built around Actors, which are programs that run on Apify's infrastructure and save structured results. Includes a store of ready-made Actors, proxies and an MCP server.
Free plan available; paid plans from USD 17 per month, billed annually
Free plan
API for searching the web, scraping pages and crawling sites, returning clean Markdown or structured data for AI applications. Offers SDKs, a CLI and an MCP server; plans are credit-based.
Free plan available; paid plans from USD 16 per month, billed annually
Free plan
Hugging Face Hub is a platform for hosting and sharing machine learning models, datasets and demo apps (Spaces) in Git-based repositories. It offers model and dataset cards, a dataset viewer, Inference Providers for serverless model calls, Inference Endpoints for dedicated deployments, Spaces hardware including ZeroGPU, and organization features such as SSO, audit logs and storage regions on Team and Enterprise plans.
Free plan available; paid plans from USD 9 per month
Free plan