Mistral OCR
The most recommended of 11 OCR models
People recommend Mistral OCR for accurate document parsing with tables, equations and bounding boxes, at low per-page cost. Critics report hallucinated sentences in some versions.
People recommend Mistral OCR for accurate document parsing with tables, equations and bounding boxes, at low per-page cost. Critics report hallucinated sentences in some versions.
109 people gave an opinion
- recommend it
- 69
- criticize it
- 40
- approval share
- 63%
Recommended in 12 of 13 months
- November 2025: recommended
- December 2025: recommended
- January 2026: recommended
- February 2026: recommended
- March 2026: recommended
- April 2026: recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: recommended
- August 2026: recommended
- September 2026: recommended
- October 2026: recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 69 people recommend itNeeds 10 or more, met.
- 1.7 recommenders for every criticNeeds 2 to 1 or better, not met.
- Recommended in 12 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Accurate output on documents
“The accuracy is great, bounding boxes are must have for proper grounding for building answers by LLMs.”
Comment from 13 Aug 2026 in “Mistral OCR 4.1” “Converting from PDF to markdown while retaining tables correctly seems to work well for me”
Comment from 15 Jan 2026 in “Ask HN: How are you doing RAG locally?” “I have been quite happy with Mistral OCR for the documents I needed to process”
Comment from 13 Aug 2026 in “Mistral OCR 4.1” - Comment from 14 Aug 2026 in “Mistral OCR 4.1”
- Comment from 23 Jun 2026 in “Mistral OCR 4”
- Comment from 23 Jun 2026 in “Mistral OCR 4”
- Comment from 1 Jul 2026 in “Leanstral 1.5”
Low cost per page
“it just runs for a year without any worries about the amount of requests I make, because the cost is minuscule”
Comment from 4 Jul 2026 in “Leanstral 1.5: Proof Abundance for All” “I prefer using a European provider though so this is actually cheaper than I can find elsewhere.”
Comment from 30 Oct 2026 in “We're releasing the weights of Mistral Large 4” “it's cheap (2c per page), fast, easy to integrate API, and sufficient for my needs”
Comment from 10 Jun 2026 in “Ask HN: What are tools you have made for yourself since the advent of AI” - Comment from 14 Aug 2026 in “Mistral OCR 4.1”
- Comment from 23 Jun 2026 in “Mistral OCR 4”
- Comment from 13 Aug 2026 in “Mistral OCR 4.1”
Fast processing
“Results were good, and fast... but I actually got better quality output from sending it to Haiku 4.5 instead.”
Comment from 18 Mar 2026 in “Mistral AI Releases Forge” “it's cheap (2c per page), fast, easy to integrate API, and sufficient for my needs”
Comment from 10 Jun 2026 in “Ask HN: What are tools you have made for yourself since the advent of AI” “in internal benchmarks, Mistral OCR is significantly faster than comparable APIs”
Comment from 13 Aug 2026 in “Mistral OCR 4.1” - Comment from 14 Aug 2026 in “Mistral OCR 4.1”
Data stays in EU
“I prefer using a European provider though so this is actually cheaper than I can find elsewhere.”
Comment from 30 Oct 2026 in “We're releasing the weights of Mistral Large 4” “it hits the exact sweet spot between cost and my data staying in the EU”
Comment from 18 Mar 2026 in “Mistral AI Releases Forge”
What people warn about
Each mark is one comment
Hallucinates sentences in documents
“For the most part Mistral is actually good, but I can't have it just inventing sentences in the middle of a document.”
Comment from 14 Aug 2026 in “Mistral OCR 4.1” “it caught entirely hallucinated sentences created by Mistral OCR 4.0”
Comment from 13 Aug 2026 in “Mistral OCR 4.1” “the hallucinations in 4.0 were so egregious (just completely making up new sentences in the middle of a page)”
Comment from 14 Aug 2026 in “Mistral OCR 4.1” Claude or Haiku gave better output
“Results were good, and fast... but I actually got better quality output from sending it to Haiku 4.5 instead.”
Comment from 18 Mar 2026 in “Mistral AI Releases Forge” “The quality I was getting from Mistral OCR 2 was nowhere near as good as what I could get from just sending the same files to Claude Sonnet”
Comment from 18 Mar 2026 in “Mistral AI Releases Forge” “found it was better just to write a program that sent the documents to Claude Sonnet to OCR instead”
Comment from 12 Apr 2026 in “European AI. A playbook to own it. By Mistral” Confusing portal and documentation
“the Mistral documentation and dashboards are all over the place”
Comment from 8 Sep 2026 in “Mistral raises €3B” “the Mistral portal is honestly kind of confusing, considering that most people just want to enter CC details and get an API key”
Comment from 29 Oct 2026 in “We're releasing the weights of Mistral Large 4” “It's even more confusing when you want to create an API key because they are separated by product”
Comment from 18 Mar 2026 in “Mistral AI Releases Forge” Accuracy below open-source alternatives
“I don’t know how they can make this statement with 79% accuracy rate. For any serious use case, this is an unacceptable number.”
Comment from 19 Dec 2025 in “Mistral OCR 3” “I'm reading worse performance than many OSS offerings like Paddle, MinerU, MonkeyOCR, etc”
Comment from 19 Dec 2025 in “Mistral OCR 3”
Other OCR models people recommend
The four most recommended other OCR models, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Mistral OCRThis page | 69 | 40 | 63% | 12 | — | — |
| PaddleOCR | 19 | 4 | 83% | 8 | — | Approved |
| GLM-OCR | 16 | 4 | 80% | 7 | — | Approved |
| DeepSeek-OCR | 12 | 4 | 75% | 9 | — | Approved |
| dots.ocr | 8 | 1 | 89% | 5 | — | — |
What people said
40 of the 127 comments that mention it, newest first. Each links to the original comment.
“I prefer using a European provider though so this is actually cheaper than I can find elsewhere.”
“the Mistral portal is honestly kind of confusing, considering that most people just want to enter CC details and get an API key”
“the Mistral documentation and dashboards are all over the place”
“Mistral OCR is really good and their small models are great for simple translations”
“it was a champ: Reasonable Markdown format with accurate content”
“For the most part Mistral is actually good, but I can't have it just inventing sentences in the middle of a document.”
“it's MUCH cheaper and faster and does an excellent job on simple ones”
“the hallucinations in 4.0 were so egregious (just completely making up new sentences in the middle of a page)”
“It's also cheaper than the murican ones from Google, Amazon”
“it caught entirely hallucinated sentences created by Mistral OCR 4.0”
“It's so bad at my work that I didn't register it might be the best in its category.”
“I have been quite happy with Mistral OCR for the documents I needed to process”
“You can even run it on your desk if you want, a single gtx 4090 is enough. It can be fully air gapped.”
“The accuracy is great, bounding boxes are must have for proper grounding for building answers by LLMs.”
“in internal benchmarks, Mistral OCR is significantly faster than comparable APIs”
“OCR is a commodity at this point, and open source models, AWS, etc already do it out of the box.”
“it just runs for a year without any worries about the amount of requests I make, because the cost is minuscule”
“OCR is off the charts good on every metric you can think of.”
“the few samples I tried showed some real improvements since dec 2025 version”
“normal handwriting got accurate but a slight different style got detected as kannada”
“Mistral has been highly accurate (their price is amazing, too).”
“If you have an OCR problem, Mistral OCR 4 is probably great.”
“Mistral OCR is able to do things no "traditional" OCR was ever able to.”
“it's cheap (2c per page), fast, easy to integrate API, and sufficient for my needs”
“documentation that is so concise, well laid out, and immediately actionable”
“found it was better just to write a program that sent the documents to Claude Sonnet to OCR instead”
“Results were good, and fast... but I actually got better quality output from sending it to Haiku 4.5 instead.”
“It's even more confusing when you want to create an API key because they are separated by product”
“it hits the exact sweet spot between cost and my data staying in the EU”
“The quality I was getting from Mistral OCR 2 was nowhere near as good as what I could get from just sending the same files to Claude Sonnet”
“Mistral lets you use both Vibe and their API through your vibe subscription api key which is incredible”
“did really good handwriting transcription, considering how tight and small some of the handwriting is”
“all-in-one solutions like Mathpix and Mistral-OCR which are quite good for my purpose”
“Converting from PDF to markdown while retaining tables correctly seems to work well for me”
“I am using the mistral API for personal projects through vim-ai and liteLLM.”
“Of course currently Mistral has an insane free tier, 1 billion tokens for each(?) of their models per month.”
“I don’t know how they can make this statement with 79% accuracy rate. For any serious use case, this is an unacceptable number.”
“I'm reading worse performance than many OSS offerings like Paddle, MinerU, MonkeyOCR, etc”
“MistralOCR works stably for me when first uploading the file to their server”
“Mistral OCR: Unmatched, the only tool that parsed our equations correctly”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.