ThermEval - Modality Identification (LLVIP): leaderboard
Metric: Accuracy (times 100, 0-100) on Task 1, thermal or RGB, on balanced thermal and RGB images of LLVIP elevated street views (5,000); zero-shot with fixed one-word or one-number prompts, a single forward pass; Gemini 2.5 models parse free-form answers into the label or number only when the output deviates from the format; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3-8B | 100 | #606 |
| 2 | InternVL3-38B | 100 | #395 |
| 3 | Qwen 2.5 VL 7B Instruct | 100 | #643 |
| 4 | InternVL3-14B | 99 | #494 |
| 5 | Qwen 2.5 VL 32B Instruct | 99 | #443 |
| 6 | Qwen 2 VL 7B Instruct | 98 | #816 |
| 7 | MiniCPM-V-2.6 | 96 | #825 |
| 8 | Llama 3.2 11B Instruct | 92 | #1112 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=thermeval-modality-identification-llvip · How It Works · Data refreshed daily, snapshot 2026-10-11.