Mbzuai released k2 horizon, six fully open models up to 375 billion parameters under apache 2.0,…
mbzuai released k2 horizon, six fully open models up to 375 billion parameters under apache 2.0, with a 512k context window and roughly 23 billion active parameters per token. the open-weight field now has a family that can be audited, modified, and deployed in pieces.
the interesting question is not the flagship, it is whether the smallest member gets more real use.
Context
MBZUAI's release of 3 September 2026 launches K2 Horizon, six models from 0.9B to 375B parameters, with models and code under Apache 2.0. The flagship 375B model has 23B active parameters in a sparse mixture-of-experts, a 36B model has 4B active, and the 375B model card says native 524,288-token (512K) context from the midtraining stages onward.
The specifications match the first-party sources. The 512K is confirmed for the flagship only, not each of the six. Fully open is MBZUAI's wording, and the exact list of released artifacts was not itemized in the part read. The largest fully open release in AI history is MBZUAI's own superlative, not independently verified. Performance claims are vendor claims, and the model card's benchmark table mixes harnesses, so scores should not be merged across rows or models. Whether the smallest member gets real use has no usage data.
Related work
- K2 Horizon project page ↗Snippet only.
- IFM releases K2 Horizon (NYU Shanghai RITS) ↗Secondary, snippet only.
Watch next
- Independent evaluations of the 0.9B to 7B models.
Sources
- K2 Horizon launch (MBZUAI, 3 Sep 2026)mbzuai.ac.ae
- K2-Horizon-375B-A23B (Hugging Face)huggingface.co
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 18 September 2026 at 19:35 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →