Mistral OCR API by Mistral AI

# Fix list: Mistral OCR API

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/mistral-ocr, the October 2026 research run, assessed 1 October 2026. Grade C, 59 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Mistral OCR API: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Security & auth, 35 out of 100, up to 11.4 more on the total

Why it scored 35: Model reading, scored like the other Mistral listings. Plain workspace API keys, revocable, no endpoint scopes (20 of 30). The same key reaches files, fine-tuning, agents and batch jobs, so it can't be limited to OCR (10 of 20). OCR returns untrusted document text, and we found no prompt-injection guidance (0 of 15). No per-call log found (0 of 15). security.txt valid per the provenance check, no certification or bug bounty confirmed this run (5 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 2. Reliability, 45 out of 100, up to 11 more on the total

Why it scored 45: status.mistral.ai on Rootly has an OCR API component with 90-day uptime bars (20). That component reads 99.31 per cent over 90 days, about 15 hours lost, and the history lists five OCR incidents since 4 August. An availability drop for Mistral OCR 4 on 21 September lasted 2 hours 42 minutes, mistral-ocr-2512 was unavailable on 4 September and OCR 4.1 was degraded the same night, and the OCR API was degraded on 4 and 25 August. Several majors (0). Limits are set per workspace tier in the console and we found no published numbers for OCR, matching the other Mistral listings (5 of 15). The error glossary says how to resolve each status code, no Retry-After confirmed (10 of 15). No SLA found (0). OCR 4.1 has been GA since 31 August 2026 (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 3. Payments & pricing, 40 out of 100, up to 7.5 more on the total

Why it scored 40: No x402, MPP or L402 (0). $4 per 1,000 pages for OCR 4.1 and $5 with Document AI annotations, on the public pricing page (20). Free Experiment tier with no card, though it needs a phone number and its data may be used for training, per the other Mistral listings (20). Browser sign-up and phone verification (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 4. Maintenance & community, 48 out of 100, up to 4.6 more on the total

Why it scored 48: Model reading. OCR 4.1 went GA on 31 August 2026, 31 days ago, and the July 2026 entry added block granularity (20 of 30). Churn is high. OCR 4.0 arrived on 23 June 2026 and retired on 30 September, about three months, per the listing's dates (5 of 20). Dated changelog and console support, with the client-python issue replies the embeddings listing found unanswered (8 of 15). Official SDKs, mistralai on PyPI and @mistralai/mistralai on npm, release dates not checked (10 of 15). SDKs generated from the OpenAPI spec, CI not checked (5 of 10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## 5. Agent ergonomics, 83 out of 100, up to 2.8 more on the total

Why it scored 83: Model reading adapted to OCR. Responses can be cut to selected pages, images are left out unless include_image_base64 is set, and blocks and confidence scores are opt-in (22 of 25). pages, include_blocks, extract_header and extract_footer and the confidence granularity are the output controls (16 of 20). Error glossary with a fix per status (15 of 20). A synchronous, stateless call that's safe to retry, with no retry guidance confirmed (15 of 20). Two required fields, model and document, and official SDKs in Python and TypeScript (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 6. Transparency & trust, 77 out of 100, up to 2 more on the total

Made of editorial 57, provenance 96.

Why it scored 77: Closed models under commercial terms with a French legal entity, SDKs Apache-2.0 (15). Abuse logs kept 30 days unless zero retention is bought, free-tier data may train models, and the paid default isn't spelt out (15 of 30). The lifecycle page promises 6 months' notice for GA models and 1 month for Labs and preview, but OCR 4.0's three-month life doesn't fit the GA period and we couldn't find its stage or announcement date (12 of 20). EU hosting by default with opt-in regional endpoints and a published subprocessor list, per the moderation listing's check (15 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: mistral.ai, registered 2019-05-15 (7 years) (11 of 15)

## 7. Schema & documentation, 89 out of 100, up to 1.8 more on the total

Why it scored 89: Model reading. Public OpenAPI document at docs.mistral.ai/openapi.yaml covering /v1/ocr (25). llms.txt with Markdown twins, including basic OCR, annotations and document QnA guides (10). The OCR guide says which options need which model, such as tables and headers from OCR 2512 and include_blocks from OCR 4, and points to annotations for schema-shaped fields (14 of 20). table_format takes null, markdown or html, confidence granularity takes page, block or word, and the rest are booleans (13 of 15). Code examples on each guide and the shared error glossary (12 of 15). Dated model ids and a dated changelog (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- Whether OCR 4.0 was GA or preview, and when its 30 September retirement was announced. A GA model retired three months after release would break the 6-month notice policy
- unchecked: durations of the 4 August, 25 August and 4 September OCR incidents
- unchecked: OCR rate limits per tier and any batch discount for OCR
- The listing's lastRelease of 2026-07-16 predates the 31 August GA of OCR 4.1. Patched, and the Models detail now says OCR 4.0 has retired

## Weaknesses

- OCR API at 99.31 per cent over 90 days, with a 2 hour 42 minute OCR 4 availability drop on 21 September
- OCR 4.0 lasted about three months before retiring on 30 September
- Free tier data may be used for training
- Workspace keys can't be limited to OCR
- No official MCP server

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Pin `mistral-ocr-4-1` rather than `mistral-ocr-latest` if output format matters downstream
- Set `table_format` to `html` for tables with merged cells
- Leave `include_image_base64` off unless you need the images, it inflates the response
- Pass `pages` to OCR only the pages you need, since billing is per page
- Upload private files through `/v1/files` and pass the signed URL

## What the review panel asked for

- Document Retry-After
- longer model lifetimes

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.