Samtool Supported Models -

Originally developed for edge computing scenarios, Samtool has gained traction in data center environments as well. Its primary value proposition lies in its ability to take a standard model format and automatically apply hardware-specific optimizations—such as operator fusion, quantization, memory planning, and kernel autotuning—without requiring developers to rewrite their models for each target device.

| Model ID | Architecture | Backbone | Output Mask Quality | Speed | Use Case | | :--- | :--- | :--- | :--- | :--- | :--- | | sam_vit_b | ViT-Base | 80M params | High | Fast | Edge devices, real-time apps | | sam_vit_l | ViT-Large | 300M params | Higher | Medium | General purpose | | sam_vit_h | ViT-Huge | 636M params | Highest | Slow | High-precision annotation | samtool supported models

If you are instead looking for the bioinformatics toolkit SAMtools , it does not use "models" in the AI sense. It is a suite of utilities for processing sequence alignment files in SAM , BAM , and CRAM formats. It is frequently used alongside BCFtools for variant calling. Introducing Meta Segment Anything Model 2 (SAM 2) It is a suite of utilities for processing

If you don’t see your specific model above, follow this 3-step check: Originally developed for edge computing scenarios

For somatic mutation calling, SAMtools provides samtools mpileup with the -B (disable BAQ) flag for tumor samples to avoid over-filtering, followed by with the -c (consensus caller) or somatic-specific priors.