Zhipu released GLM-5.2, achieving stable long-horizon task capability in a 1M-token context for the first time. Its IndexShare architecture shares a single indexer across every four sparse attention layers, reducing FLOPs per token by 2.9x at a 1M context. The model is open-sourced under the MIT license with no regional restrictions.
Key Takeaways
Zhipu released GLM-5.
2,achieving stable long-horizon task capability in a 1M-token context for the first time.
Its IndexShare architecture shares a single indexer across every four sparse attention…
9x at a 1M context.
The model is open-sourced under the MIT license with no regional restrictions.
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.