Dimensions overall score 9.0

GitHub Code Pulse

No public code linked for this paper yet.

• Surg-R1 is a surgical Vision-Language Model designed to provide interpretable reasoning for surgical scene understanding.technical 95%
• Surg-R1 utilizes a three-level reasoning hierarchy: perceptual grounding, relational understanding, and contextual reasoning.method 95%
• The model was trained using a four-stage pipeline.method 95%
• A surgical chain-of-thought dataset with 320,000 reasoning pairs was created for training Surg-R1.method 95%
• Surg-R1 achieves an Arena Score of 64.9% on public benchmarks from SurgBench.result 95%
• Surg-R1 outperforms Gemini 3.0 Pro (46.1%) and GPT-5.1 (37.9%) on SurgBench public benchmarks.result 95%
• Surg-R1 demonstrates a 15.2 percentage point improvement over the strongest surgical baseline on external validation datasets.result 95%
• Surg-R1 is a Vision-Language Model designed for surgical decision support that incorporates hierarchical reasoning.technical 95%
• Surg-R1 employs a three-level reasoning hierarchy: perceptual grounding, relational understanding, and contextual reasoning.method 95%
• The largest surgical chain-of-thought dataset, containing 320,000 reasoning pairs, was created for training Surg-R1.method 95%
• Surg-R1 was trained using a four-stage pipeline, progressing from supervised fine-tuning to group relative policy optimization and iterative self-improvement.method 95%
• Surg-R1 achieved the highest Arena Score (64.9%) on public benchmarks compared to Gemini 3.0 Pro (46.1%) and GPT-5.1 (37.9%).result 95%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card