ParseBench - Visual Grounding: leaderboard
LlamaIndex's ParseBench scores document parsing of human-verified enterprise pages from insurance, finance and government documents with rule-based checks; VLM rows only, parser products excluded. This board reports the Visual Grounding dimension: element pass rate for tracing extracted elements to their location on the page. Models that emit no layout boxes (upstream score exactly 0) are omitted.
Metric: Element Pass Rate. Source: parsebench.ai. Saturation forecast: Around October 2028. 112 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 70.85 | #66 |
| 2 | Gemini 3.8 Flash (High) | 69.03 | #30 (Gemini 3.8 Flash) |
| 3 | Gemini 3.8 Flash (Low) | 67.81 | #30 (Gemini 3.8 Flash) |
| 4 | Gemini 3.8 Flash (Medium) | 66.79 | #30 (Gemini 3.8 Flash) |
| 5 | Gemini 3.7 Flash (High) | 66.4 | #24 (Gemini 3.7 Flash) |
| 6 | Gemini 3.7 Flash (Low) | 66.31 | #24 (Gemini 3.7 Flash) |
| 7 | Gemini 3.1 Pro (Preview) (Low) | 64.69 | #66 (Gemini 3.1 Pro (Preview)) |
| 8 | Gemini 3.7 Flash (Medium) | 64.39 | #24 (Gemini 3.7 Flash) |
| 9 | Gemini 3.5 Flash (High) | 63.95 | #68 (Gemini 3.5 Flash) |
| 10 | Claude Fable 5.1 | 63.81 | #10 |
| 11 | Gemini 3.6 Flash (Minimal) | 62.72 | #59 (Gemini 3.6 Flash) |
| 12 | Gemini 3.6 Flash (Medium) | 62.5 | #59 (Gemini 3.6 Flash) |
| 13 | Gemini 3 Flash (High) | 62.37 | #105 (Gemini 3 Flash) |
| 14 | Gemini 3 Flash (Medium) | 61.53 | #105 (Gemini 3 Flash) |
| 15 | Gemini 3.5 Flash (Medium) | 61.15 | #68 (Gemini 3.5 Flash) |
No result here: #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=parsebench-visual-grounding · How It Works · Data refreshed daily, snapshot 2026-11-07.