RadLE 2.0 (Radiology): leaderboard
RadLE 2.0 (crashlab.in) — radiology visual-reasoning cases pitting vision-language models against board-certified radiologists and trainees; models earn an Elo-style RadLE score (human experts sit near 989).
Metric: RadLE Score. Source: crashlab.in. Saturation forecast: Around October 2027. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 | 758 | #16 |
| 2 | Muse Spark 1.1 | 735 | #52 |
| 3 | GPT-5.6 Pro Sol | 665 | #18 |
| 4 | Gemini 3.1 Pro (Preview) | 660 | #58 |
| 5 | Grok 4.5 | 524 | #48 |
| 6 | Qwen 3.7 Plus | 472 | #92 |
| 7 | GLM-5V Turbo | 409 | #166 |
| 8 | MiniMax-M3 | 350 | #136 |
| 9 | Gemma 4 31B | 336 | #200 |
| 10 | MedGemma 1.5 4B | 277 | #789 |
| 11 | Llama 4 Maverick | 263 | #464 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=radle-2-0-radiology · How It Works · Data refreshed daily, snapshot 2026-10-20.