Blueprint-BenchBrowse 284 families
Blueprint-Bench
Can the model extract and reason over information in images or documents?
Version / measure
- Domain
- Vision
- Ranked models
- 8
- Reported values
- 9
- Reports
- 2 · lab
- Direction
- Higher is better
- Status
- Version not specified
Rankings
Best reported score per model across effort settings
| # | Model | Best setting | Score |
|---|---|---|---|
| 1 | GPT-5.5OpenAI | reportedcost n/a2 values · Jun 9 · source | 36.2% |
| 2 | Gemini 3.5 FlashGoogle | reportedcost n/a1 value · May 19 · source | 33.6% |
| 3 | Gemini 3.1 Pro PreviewGoogle | reportedcost n/a1 value · Jun 9 · source | 26.5% |
| 4 | Gemini 3.1 ProGoogle | reportedcost n/a1 value · May 19 · source | 26.5% |
| 5 | Claude Opus 4.7Anthropic | reportedcost n/a1 value · May 19 · source | 24.5% |
| 6 | Claude Opus 4.8Anthropic | reportedcost n/a1 value · Jun 9 · source | 14.5% |
| 7 | Claude Sonnet 4.6Anthropic | reportedcost n/a1 value · May 19 · source | 6.7% |
| 8 | Gemini 3 FlashGoogle | reportedcost n/a1 value · May 19 · source | 0.0% |
Effort curve
Score vs. reported cost. Lines connect complete sweeps only.
No cost-linked effort sweep for this version.
Definition and comparison boundarypublic methodology
Can the model extract and reason over information in images or documents?
Visual perception and multimodal reasoning under the reported protocol. Higher is better. The value is the percentage reported in this lab's table.
Harness, tools, prompts, attempt count, and benchmark version can materially change the result. Compare within one reporting context.
Normalized score.
Method / source ↗