Article · 2026-06-28 · 6 moments

Latest open artifacts (#22): Zyphra, Cohere, and Poolside are expanding the breadth of the ecosystem

An assessment of the open ecosystem and the motivations behind releasing models ✦ AI generated

Prediction

As the open model ecosystem diversifies, more companies will build a long tail of specialized models, while fewer companies will keep chasing the absolute, bleeding-edge open frontier.

The author predicts that the growing diversity of model makers will produce a long tail of niche, specialized models even as fewer companies compete for the very top of open-source performance.

transcript

the author: This diversity of makers and models fits our hypothesis that more companies will develop a long-tail of models and the number of companies chasing the absolute, open frontier will diminish.

gives example · 3provides context · 1

Data

Cohere released its flagship Command A+ model under the permissive Apache 2.0 license, a major change from prior non-commercial licensing, as a 218B-A25B MoE that can run on a single B200 GPU at 4-bit precision.

Cohere shifted its flagship Command A+ model to an Apache 2.0 license after previously using non-commercial terms, releasing a 218B-A25B mixture-of-experts model runnable on a single B200 GPU.

transcript

the author: Cohere, which is becoming more of a regular entrant into Artifacts lately, released their flagship, Command A+, under Apache 2.0. Previous iterations of the series have been released under a non-commercial license, so this change is more than welcome! Command A+ combines multi-modal, multi-lingual and agentic capabilities as a 218B-A25B MoE, making it usable with a single B200 (when using 4-bit).

Claim◆

Open weights are now Poolside's default approach, and the company will keep building toward the frontier while releasing increasingly capable models openly.

After releasing its flagship Laguna-M.1 model under Apache 2.0, Poolside publicly committed to open weights as its default going forward for future, increasingly capable models.

transcript

Poolside: Poolside, which we covered in the last Artifacts, also released their flagship model under Apache 2.0! They also commit to open releases going forward: Open weights are now our default. We’ll keep building toward the frontier and releasing increasingly capable models in the open.

Context

The open model landscape has shifted from being dominated by a handful of Chinese players a year ago to now featuring a much wider, more diverse set of niche companies from all over the world.

The author observes that the open-source AI model ecosystem has become far more diverse in the past year, expanding well beyond the Chinese labs that once dominated it.

transcript

the author: A year ago, open artifacts and the open model landscape more broadly were dominated by a handful of (Chinese) players. This has shifted, with us increasingly featuring more niche companies all over the world.

gives example · 2

Claim◆

Trying to slow or ban the open model ecosystem is futile, unsafe, and anti-freedom, because it would concentrate AI development among a select few and cut off outsiders' ability to adopt the technology.

The author argues that policy efforts to restrict open-weight AI models would fail like past tech bans, while also being unsafe and freedom-restricting because they would concentrate AI power in a few hands.

transcript

the author: Attempts to slow or ban this ecosystem are not only futile, as the history of tech-related bans has shown, but also unsafe and anti-freedom. Such restrictions would concentrate AI development and usage among the select few, which ultimately endangers outsiders’ ability to freely adopt one of the most important technologies of our lifetime.

explains mechanism · 1provides context · 1supports · 6

Fact

Zyphra, which trains its models on AMD GPUs and is seen as an insider tip in the research community for its detailed architecture-focused tech reports, released new flagship MoE models: a 74B-A4B and an 8B-A0.6B.

Zyphra, known for training on AMD hardware and publishing architecture-rich tech reports, released two new flagship mixture-of-experts models, a 74B-A4B and an 8B-A0.6B.

transcript

the author: Zyphra, which trains on AMD GPUs and is known as some sort of insider tip in the research community due to their tech reports with interesting architecture choices, has released some new models, with a 74B-A4B MoE and an 8B-A0.6B MoE (tech report) being their current flagship releases.

Highlight slides

Banning Open Models Is Futile✦ from: Trying to slow or ban the open model ecosystem is futile, unsafe, and anti-freedom, because it would concentrate AI development among a select few and cut off outsiders' ability to adopt the technology.Poolside Makes Open Weights the Default✦ from: Open weights are now Poolside's default approach, and the company will keep building toward the frontier while releasing increasingly capable models openly.Concentration Endangers Access✦ from: Trying to slow or ban the open model ecosystem is futile, unsafe, and anti-freedom, because it would concentrate AI development among a select few and cut off outsiders' ability to adopt the technology.Open Source, Not Open Ceiling✦ from: Open weights are now Poolside's default approach, and the company will keep building toward the frontier while releasing increasingly capable models openly.

Related episodes

Article · Latent Space

[AINews] OpenAI GPT-5.6 Sol / Terra / Luna — restricted to trusted partners

→

Video · No Priors

How Nuclear Will Unlock Energy Abundance with Valar Atomics Founder Isaiah Taylor

→

Article · Import AI

Import AI 462: Superpersuasion; self-sustaining AI; paths to ASI

→