rankings / read-documents
The best models for reading documents
96 models & services · 0 callable here now
Ranked by benchmark score per dollar (quality floor applied). Scores: olmOCR-bench — how accurately a model reads and parses real documents. Fetched 2026-07-04. Prices are our live per-call rates; ~ marks an estimate until the model is onboarded.
sorted by value · sort by score
| # | model | score | price | params | status |
|---|---|---|---|---|---|
| 1 | tiiuae/Falcon-OCR best value | 80.3 (#8) | ~$0.094 / 1k images | 269.9M | coming soon |
| 2 | datalab-to/surya-ocr-2 | 83.3 (#4) | ~$0.235 / 1k images | 686.2M | coming soon |
| 3 | lightonai/LightOnOCR-2-1B | 83.2 (#5) | ~$0.235 / 1k images | 1005.6M | coming soon |
| 4 | lightonai/LightOnOCR-1B-1025 | 76.1 (#12) | ~$0.235 / 1k images | 1161.2M | coming soon |
| 5 | opendatalab/MinerU2.5-2509-1.2B | 75.2 (#14) | ~$0.235 / 1k images | 1156M | coming soon |
| 6 | rednote-hilab/dots.mocr | 83.9 (#3) | ~$0.626 / 1k images | 3039.2M | coming soon |
| 7 | rednote-hilab/dots.ocr | 79.1 (#10) | ~$0.626 / 1k images | 3039.2M | coming soon |
| 8 | deepseek-ai/DeepSeek-OCR-2 | 76.3 (#11) | ~$0.626 / 1k images | 3389.1M | coming soon |
| 9 | datalab-to/chandra-ocr-2 | 85.8 (#2) | ~$1.341 / 1k images | 5295.6M | coming soon |
| 10 | datalab-to/chandra | 83.1 (#6) | ~$1.341 / 1k images | 8767.1M | coming soon |
| 11 | baidu/Qianfan-OCR | 79.8 (#9) | ~$1.341 / 1k images | 4741.4M | coming soon |
| 12 | infly/Infinity-Parser2-Pro top score | 87.6 (#1) | - | 35107.2M | not hosted |
| 13 | infly/Infinity-Parser-7B | 82.5 (#7) | - | 8292.2M | coming soon |
| 14 | deepseek-ai/DeepSeek-OCR | 75.7 (#13) | - | 3336.1M | not hosted |
| 15 | zai-org/GLM-OCR | 75.2 (#15) | - | 1325.3M | not hosted |
| 16 | Mistral OCR | 72src (#16) | $1.00 / 1k pages | - | reference |
| 17 | OpenAI GPT-4o | 69.9src (#17) | per-token vision pricing | - | reference |
| 18 | Google Gemini Flash 2 | 63.8src (#18) | per-token vision pricing | - | reference |
| 19 | Qwen/Qwen2.5-VL-7B-Instruct | - | ~$1.341 / 1k images | 8292.2M | coming soon |
| 20 | Qwen/Qwen3.6-35B-A3B-FP8 | - | ~$1.341 / 1k images | 35953.9M | coming soon |
| 21 | Qwen/Qwen2.5-VL-3B-Instruct | - | ~$0.626 / 1k images | 3754.6M | coming soon |
| 22 | cyankiwi/gemma-4-26B-A4B-it-AWQ-4bit | - | ~$1.341 / 1k images | 26554.3M | coming soon |
| 23 | Qwen/Qwen3.6-27B-FP8 | - | ~$1.341 / 1k images | 27782.9M | coming soon |
| 24 | Qwen/Qwen3-VL-4B-Instruct | - | ~$1.341 / 1k images | 4437.8M | coming soon |
| 25 | Qwen/Qwen2-VL-2B-Instruct | - | ~$0.626 / 1k images | 2209M | coming soon |
| 26 | llava-hf/llava-1.5-7b-hf | - | ~$1.341 / 1k images | 7063.4M | coming soon |
| 27 | RedHatAI/gemma-4-31B-it-FP8-block | - | ~$1.341 / 1k images | 31274.9M | coming soon |
| 28 | HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive | - | - | - | coming soon |
| 29 | microsoft/Florence-2-base | - | ~$0.094 / 1k images | 231.6M | coming soon |
| 30 | Qwen/Qwen3.5-0.8B | - | ~$0.235 / 1k images | 873.4M | coming soon |
| 31 | Qwen/Qwen3-VL-2B-Instruct | - | ~$0.626 / 1k images | 2127.5M | coming soon |
| 32 | RedHatAI/gemma-4-26B-A4B-it-FP8-Dynamic | - | ~$1.341 / 1k images | 26560.9M | coming soon |
| 33 | Salesforce/blip-image-captioning-base | - | - | - | coming soon |
| 34 | cyankiwi/Qwen3.6-35B-A3B-AWQ-4bit | - | ~$1.341 / 1k images | 35951.8M | coming soon |
| 35 | Qwen/Qwen2-VL-7B-Instruct | - | ~$1.341 / 1k images | 8291.4M | coming soon |
| 36 | Qwen/Qwen2-VL-7B-Instruct-AWQ | - | ~$1.341 / 1k images | 8291.4M | coming soon |
| 37 | unsloth/Qwen3.6-27B-MTP-GGUF | - | - | - | coming soon |
| 38 | Qwen/Qwen2.5-VL-7B-Instruct-AWQ | - | ~$1.341 / 1k images | 8292.2M | coming soon |