Mistral OCR 4.1
Mistral OCR 4.1 is the live latest OCR row on the Mistral catalog, not an older OCR 4.0 leftover. Rechecked 2026-08-18: docs.mistral.ai/models still features OCR 4.1 with paragraph boxes, structural labels, and confidence. mistral.ai/pricing still prints OCR 4.1 / 4.0 / 3 at $4 per 1,000 pages and OCR 2 at $1 / 1,000 pages. OCR and OCR 2 are marked deprecated. Do not invent a leftover OCR 4.0 page or a cheaper 4.1 dollar row.
Compare leftover Pixtral 12B if you wanted the leftover 2024 VLM, leftover Mistral Small if you wanted the leftover Small chat snapshot, and GPT-5.6 Luna if you wandered here for a current cheap chat model.
Key Features
- Live latest OCR: First-party catalog still features OCR 4.1.
- Printed extras: Paragraph boxes, structural labels, and confidence stay on the models page we opened.
- Same $4 / 1,000 pages row as OCR 4.0 / 3: Rechecked on the public pricing table.
- Older OCR / OCR 2 are deprecated: Do not treat them as the 2026 default.
Use Cases
- People who need the current Mistral OCR row after leftover chat / VLM pages were rewritten.
- People comparing OCR 4.1 vs deprecated OCR / OCR 2.
- People who were about to invent a cheaper 4.1 dollar row.
Limitation: This site has no leftover OCR 4.0 slug. Recheck first-party docs for the exact API id.
Pricing
First-party pages on 2026-08-18.
| Piece | Price | Notes |
|---|---|---|
| OCR 4.1 | $4 / 1,000 pages | Live latest OCR on pricing. |
| OCR 4.0 / OCR 3 | $4 / 1,000 pages | Still listed beside 4.1. |
| OCR 2 | $1 / 1,000 pages | Deprecated, cheaper row. |
If a leftover blog still treats OCR 2 as the current default, treat the 2026 catalog as the correction.
Getting Started
- Open docs.mistral.ai/models and confirm OCR 4.1 is the featured OCR.
- Recheck mistral.ai/pricing before you quote a page rate.
- Recheck leftover Pixtral 12B if you wanted the leftover VLM, not OCR.
- Do not invent an OCR 4.0 page on this site.
Frequently Asked Questions
Is OCR 4.1 cheaper than OCR 4.0?
Not on the table we opened. Both print $4 / 1,000 pages.
Same as Pixtral 12B?
No. Pixtral 12B is a leftover 2024 VLM.
Are older OCR rows gone?
OCR and OCR 2 are marked deprecated. OCR 3 and OCR 4.0 are still listed beside 4.1.
Alternatives
- Pixtral 12B: Leftover 2024 Mistral VLM, not the live OCR row.
- Mistral Small: Leftover Small chat snapshot.
- GPT-5.6 Luna: Live cheap OpenAI sibling if you wandered here for chat.
Tips
- Call OCR 4.1 the live latest OCR row, not leftover OCR 4.0.
- Quote $4 / 1,000 pages only while first-party still prints it.
- Recheck the models page for paragraph boxes, labels, and confidence before you invent extras.
Conclusion
Mistral OCR 4.1 is the live latest OCR row at $4 / 1,000 pages. Start at docs.mistral.ai/models, then decide whether a leftover Mistral chat / VLM page already covers a different job.
Comments
No comments yet. Be the first to comment!
Related Tools
Mistral Voxtral TTS
docs.mistral.ai/models
Mistral Voxtral TTS: live catalog TTS row, v 26.03, CC BY-NC 4.0. Rechecked 2026-08-18. No invented dollar row.
Mistral Shieldstral 1.0
docs.mistral.ai/models
Mistral Shieldstral 1.0: live compact multimodal moderation model. Rechecked 2026-08-18. Apache 2.0. No invented dollar row.
Mistral Voxtral Mini Transcribe 2
docs.mistral.ai/models
Mistral Voxtral Mini Transcribe 2: live Premier transcription row, v 26.02. Rechecked 2026-08-18. Older 25.07 is deprecated. No invented dollar row.
Related Insights
Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.
A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.
What Locks Codex Is the Picker, Not the Models
You already pay for OpenCode Go, Grok, and Z.ai, but the Codex picker still shows mostly GPT. The community project codex-router does not teach another install ritual. It puts subscriptions you already bought back into the selector, keeps keys on the machine, and leaves native GPT alone.
Seven AI Coding CLIs, Six Months: No Matter How Strong the Model, Work Needs Supervision
Claude Code, Codex, opencode, pi, omp and DeepSeek Harness all have personalities. After six months of deep use I run a division of labor: pi for the fastest cheapest reviews, omp for complex PRs, DeepSeek Harness on V4 Flash for high-frequency low-cost review, and Claude Code, Qoder and Cursor for writing. No matter how strong the model, work needs supervision — ideally from an independent third party.