Omni-Persona - Image-to-Image: leaderboard
Image-to-Image persona matching on Omni-Persona (2026): a query face image is matched to profile images to answer personalized questions, scored by calibrated accuracy, which combines judged answers on answerable items with correct abstention on items whose persona is absent from the context. Percent; higher is better.
Metric: Calibrated accuracy (%): item-weighted accuracy over answerable items (answer judged correct by a GPT-5.4-mini judge) and absent-persona items (correct abstention by keyword matching), on the image-to-image persona matching group (query face image, profile image context), zero-shot; higher is better. Source: arxiv.org. Saturation forecast: Around 2031. 10 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 3.6 Flash | Medium | 38.5 | #81 |
| 2 | Qwen3 Omni 30B A3B Instruct | 37.2 | #930 | |
| 3 | Qwen2.5-Omni-3B | 35.9 | #3519 | |
| 4 | Gemini 3.5 Flash Lite | 35.1 | #299 | |
| 5 | gemma-4-E4B-it | 34.6 | #1539 | |
| 6 | Qwen2.5-Omni-7B | 33.8 | #1884 | |
| 7 | MiniCPM-o-4.5 | Thinking | 31.6 | #1252 |
| 8 | Gemma 4 E2B (IT) | 26.4 | #2451 | |
| 9 | Phi-4 Multimodal Instruct | 26 | #4154 | |
| 10 | Interactive-Omni-8B | 20.8 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=omni-persona-image-to-image · How It Works · Data refreshed daily, snapshot 2026-11-21.