Microsoft releases GigaPath-Flash: Efficient pathology models cut compute by 50x
Microsoft Research · rss · 2026-09-01
Microsoft has introduced GigaPath-Flash and GigaTIME-Flash, efficient iterations of its pathology foundation models. By distilling the original billion-parameter GigaPath encoder into a compact 22M-parameter ViT-S backbone, these models retain approximately 97% of predictive performance while reducing computational costs by roughly 50 times.
- GigaPath-Flash: Optimized for whole-slide analysis, it uses a ViT-S tile encoder and LongNet slide encoder, achieving competitive results on PANDA and EBRAINS benchmarks with minimal inference cost.
- GigaTIME-Flash: Designed for spatial proteomics prediction, it replaces the CNN backbone with the distilled ViT-S encoder. It is 6x faster and uses 8x less memory than the original, with improved performance on out-of-distribution data.
These efficiency gains make population-scale discovery practical, enabling researchers to run repeated analyses across large cohorts. Both models are released under the Apache 2.0 license on Hugging Face.
More from Research
- Interpretability debate: one interpretable polysemantic neuron doesn't prove all are — aran_nayebi · 2026-10-16
- Optuna creator Akiba: when 2 hyperparameters cost $20M, what's left for BO/HPO — Stefania_druga · 2026-10-16
- AI will be superhuman at science regardless, and Europe can't opt out — teortaxesTex · 2026-10-16
- SERO augments long-horizon agentic RL with executable oracle continuations — Shanghai-AI-Laboratory · 2026-10-16
- Tsinghua and Amap propose OmniVerifier-C1, a graded verifier beating binary judges — Bowei Liu · 2026-10-16
- Alibaba open-sources MAI-Gym, an all-in-one GUI agent eval and RL training platform — Tongyi-MAI · 2026-10-16