OmniSch - Symbol Detection: leaderboard
Schematic-to-netlist conversion on OmniSch (2026): from a PCB schematic image exported by an industrial EDA tool, the model generates the netlist graph of symbols, pins and nets in one zero-shot prompt, compared with reference graphs extracted automatically from the schematic source files. F1 of component symbol detection under attribute and structure matching, times 100; higher is better.
Metric: Symbol detection F1 (x100): ground-truth component symbols matched by the symbol matching score of Eq. 1 (class, name, value, structure and coarse position, no box overlap required) on the industrial-EDA (Real) schematics of the OmniSch held-out test split, the model converting each PCB schematic image into a spatially weighted netlist graph with one zero-shot prompt, predicted and reference nodes aligned by maximum-weight bipartite matching; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 9 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | OmniSch Gemini 2.5 Flash (checkpoint unspecified) | 66.36 | |
| 2 | Gemini 3.1 Pro (Preview) | 59 | #99 |
| 3 | Claude Sonnet 4.6 | 58.62 | #173 |
| 4 | Qwen 3.5 122B A10B | 55.84 | #346 |
| 5 | GPT-5.2 | 52.76 | #185 |
| 6 | Ministral 3 14B | 38.26 | #1870 |
| 7 | GPT-5 Mini | 29.97 | #364 |
| 8 | Llama 4 Maverick | 25.27 | #1110 |
| 9 | Qwen 3 VL 8B Instruct | 8.68 | #1029 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=omnisch-symbol-detection · How It Works · Data refreshed daily, snapshot 2026-11-21.