K2-Horizon: A Frontier Fleet of Open Source Language Models
The K2-Horizon paper reports a fleet of six fully open language models, from 0.9B dense to a 375B-A23B MoE, aimed at reasoning and agentic tasks.
Source-based brief · Editor review not recorded · See checks and limits ↓
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face
- Discovery linkHugging Face
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
The K2-Horizon paper reports a fleet of six fully open language models, from 0.9B dense to a 375B-A23B MoE, aimed at reasoning and agentic tasks. It introduces Mixture-of-Value Attention and softplus attention gating, and says its 36B-A4B model nearly matches a 32B dense model trained on the same data with one-seventh the active parameters. It also describes Uno diffusion adapters for decoding speedups of up to 2.2x and releases pipelines that produced roughly 10T synthetic pre-training tokens. The supplied text is a partial page capture; the full paper was not checked.
Claim explorer
Attribution, source and conditionsSource-stated points from the stored brief. Claim-level assessments have not been recorded.
Source claims not independently checked; the supplied text is a partial page capture and benchmark results were not included.
This connects a source, an attributed statement and its qualifications. It does not establish that the claim is correct.
Questions worth asking
Open questions for further reportingWhat remains unconfirmed?
Source claims not independently checked; the supplied text is a partial page capture and benchmark results were not included.
What material was available for this brief?
The recorded brief basis is fetched source page. These points are AINewsIQ paraphrases, not quotations or independent confirmations.
Six open models range from a 0.9B dense model to a 375B-A23B MoE, released with intermediate checkpoints and training code.
Source: Hugging Face ↗Stored source-based brief · key point 1 · AINewsIQ paraphraseMixture-of-Value Attention and softplus attention gating are introduced; the 36B-A4B model nearly matches a 32B dense model on the same data.
Source: Hugging Face ↗Stored source-based brief · key point 2 · AINewsIQ paraphraseUno lossless diffusion adapters are reported to speed decoding by up to 2.2x without a separate draft model.
Source: Hugging Face ↗Stored source-based brief · key point 3 · AINewsIQ paraphrasePipelines produced roughly 10T synthetic pre-training tokens, which the authors say shows synthetic data can be a major pre-training ingredient.
Source: Hugging Face ↗Stored source-based brief · key point 4 · AINewsIQ paraphraseLimits from the stored brief
Source claims not independently checked; the supplied text is a partial page capture and benchmark results were not included.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
Related archive coverage
These relationships record a citation or topic overlap, not confirmation of the finding.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
Brief history · 1 retained version
These are changes to AINewsIQ’s own publication. Earlier history is unavailable unless recorded here. Complete retained versions are available to the editor.
- Version 1 · published
Source-based brief stored or updated. This does not establish that the publisher changed its page.
Automated record
Source observations · 1 recent captures
A changed capture means the extracted text differed. It can reflect page furniture or extraction changes; it does not establish a substantive correction. Source captures remain private.
Retained capture; no earlier comparison available.
New engagement snapshots do not mean the source itself changed. Source observations and AINewsIQ brief revisions are separate records.
Why this may matter
The release emphasizes full reproducibility, including data pipelines and RL stack, which open releases rarely disclose. Architectural and synthetic-data results could inform open model development.
Limits and open questions
Source claims not independently checked; the supplied text is a partial page capture and benchmark results were not included.
Organization-specific financial, security, and operational effects require the organization's own context and review.

Join the discussion
Have a different perspective or a question about the evidence? Discussion is free with a verified email account. Every comment is reviewed before publication.
No published comments yet.
New commenters: 2 submissions per 24 hours, at least 5 minutes apart, until 3 are approved. Discussion guidelines and moderation →