Gemini Flash vs InternVL3.5
Pick Gemini Flash for cheap bulk PDF to Markdown or JSON extraction through an API, with no GPUs to run. Pick InternVL3.5 if documents must stay on your own hardware, or you want an Apache-2.0 model in a size that fits your GPU budget. The main split is hosted per-token billing versus self-hosted compute.
Gemini Flashgoogle/gemini-flash | InternVL3.5opengvlab/internvl | |
|---|---|---|
| Type | Vision model | Vision model |
| Price | 15 credits / page | 2 credits / page |
| Per 1,000 pages | $15.00 | $2.00 |
| Typical speed | 1.6 s / page | 3.0 s / page |
| Best for | Handwriting and mixed layouts | Self-hosted document understanding |
| Inputs | PDF, images | images |
| Languages | 100+ languages | Multilingual |
| Max pages per request | 1,000 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | No |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | ✓ Yes |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose Gemini Flash if
- Gemini 3 Flash scores 92.62 on OmniDocBench v1.6, an independent score close to Gemini 3 Pro at 92.91.
- You call it through the Gemini API or Vertex AI and pay per token, starting at $0.50 input and $3.00 output per 1M tokens, or $0.25 and $1.50 on Batch/Flex.
- A context window of up to 1M tokens fits long documents in one request, and it handles mixed-language documents.
Choose InternVL3.5 if
- It is Apache-2.0 and self-hosted, so it can run in air-gapped or on-prem setups where Gemini Flash cannot.
- Nine sizes from 1B to 241B let you trade accuracy for cost, and models up to 30B fit on a single A100.
- There is no per-token fee, only your compute, and the MoE variants 20B-A4B and 30B-A3B activate few parameters per token.
Key differences
| Gemini Flash | InternVL3.5 | |
|---|---|---|
| How you run it | Hosted API (Gemini API or Vertex AI) | Self-host with Transformers, LMDeploy or vLLM |
| Licence | Proprietary | Apache-2.0 |
| Price | $0.50 input / $3.00 output per 1M tokens for Gemini 3 Flash Preview; Batch/Flex $0.25 / $1.50 | Free software; you pay for compute |
| OCR benchmark | OmniDocBench v1.6 overall: 92.62 (Gemini 3 Flash) | No text scores. The model card shows OCR results only as chart images |
| Hardware and limits | No hardware to manage; 1M token context window | 30B and under on one A100; 38B needs two A100s; 241B-A28B needs eight A100s |
| Output formats | Markdown, JSON, HTML, plain text, driven by your prompt | Plain text, Markdown, JSON |
| Known weak spots | Cost per page varies with page density and output length. No fixed schema or confidence scores. Several newer Flash models rise from $0.75/$3.75 to $1.50/$7.50 per 1M tokens on January 1, 2027. | No bounding boxes or layout output by default. Larger sizes are not practical on CPU only. |
Sources
- Gemini API pricing
- OmniDocBench leaderboard
- Gemini pricing changes 2026
- Hugging Face model card (8B)
- InternVL GitHub
What each one is
Gemini Flash
General vision model with strong handwriting and layout understanding at a low price.
Gemini Flash detailsInternVL3.5
Shanghai AI Lab's open vision-language family (1B to 241B, August 2025), strong on OCR and document benchmarks.
InternVL3.5 detailsTry both on your own file
Run the file once with each model and compare the output.
PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?
Use one, fall back to the other
{"model": "google/gemini-flash","fallbacks": ["opengvlab/internvl"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Gemini Flash vs MiniCPM-V 4.5Gemini Flash vs PixtralGPT vs InternVL3.5Mistral OCR vs InternVL3.5Gemini Flash vs Azure Document IntelligenceGemini Flash vs Granite DoclingSmolDocling vs InternVL3.5Surya vs InternVL3.5
Gemini Flash alternativesInternVL3.5 alternativesBest OCR models in 2026