K2 Horizon
Star of the week, issue #11What it is
A 'connected fleet' of six Apache-2.0 models from 0.9B to 375B parameters, out of the Institute of Foundation Models at Abu Dhabi's MBZUAI — the LLM360 lineage. The flagship is a 375B MoE with 23B active and 512K context; the 0.9B targets watches and glasses, the 3.7B and 7B target phones, and those small models ship the full stack — weights, training data, recipe, and code. Runs on vLLM, SGLang, and Ollama.
Featured in
Issue #11 — Mark Your Own Homework · — the full issue has that week's other tools, reads, and big-lab releases.
More models tools we've covered
- Strands Decider 2B A small open model that makes typed decisions — pick one, yes or no, rate it — with a confidence score, in one forward pass.
- Laya An open-weight 'System 1' decision engine: you pass it any state — an email, a ticket, a JSON document — plus typed questions (pick one of these, score this, yes or no), and it answers all of them in a single forward pass with calibrated confidence and no text generation at all. Under the hood are three small encoder checkpoints (ModernBERT-large for English, a 322M-parameter mmBERT for 100+ languages, and a fine-tuned one), with a router that picks per request. `pip install laya`, with optional extras for an MCP server, an HTTP server, LangChain, and ONNX. Apache-2.0.
- Jev TypeSafe AI's 'System One model' generates no text at all: you hand it structured program state and up to 255 options, and it returns a typed decision — classify, route, score, extract — with calibrated confidence, in 70 to 500 milliseconds. The pitch is replacing the small LLM calls buried inside pipelines, where parsing prose was always the bug. Pricing is $0.042 per million input tokens with output free; the catch is that it's a closed API behind an early-access waitlist.