Open AI Models: New Wave Closes Gap with Closed Systems
A new wave of five major open AI models launched in July 2026, significantly narrowing the performance gap with closed, proprietary systems. This development marks a pivotal shift in the AI landscape.
What happened?
Within a two-week period in July 2026, five significant open AI models were released. These include Moonshot's Kimi K3 with 2.8 trillion parameters, Thinking Machines' Inkling under the Apache 2.0 licence, and DeepSeek's official V4 release. Additionally, news emerged regarding MiniMax 2.7T 'M3 Pro' and Mistral's new family of sparse MoE (Mixture of Experts) models opening for early access.
Key facts
”Five open-weight moves landed in one July window — Kimi K3 shipped, Thinking Machines put the US back in the race with Inkling, MiniMax and Mistral teased much larger MoEs, and DeepSeek’s official V4 sits on a mid-July schedule.”
”The open-vs-closed gap now reads as roughly one model generation, not several.”
Why it matters
This concentrated wave of launches demonstrates that the gap between open and closed AI models now corresponds to approximately one model generation, rather than several. This implies that access to advanced AI technology is being democratised, which could accelerate innovation and allow more actors to develop AI applications without being dependent on a few large technology companies. The development of open models, such as Mistral's MoE, also signals increased complexity and capacity within the field.
Who is affected?
Developers and researchers within the AI sector are directly affected, gaining access to more powerful tools for experimentation and product development. Companies building on AI models can also benefit from increased flexibility and reduced costs. Users of AI products can expect faster innovation and more tailored solutions in the long term.
Impact on the EU
The launch of these open-source models is global in nature, and their availability positively impacts EU-based developers and companies. Although individual models may have different licensing structures, the general trend of 'open-weight' releases promotes broader access within the EU.
What else you should know
K3's open weights are expected to become available on 27 July 2026. Inkling features 975 billion parameters, of which 41 billion are active, and is licensed under Apache 2.0. Analysis by Digital Applied connects these five events, which are otherwise not covered collectively by individual providers.
Quick answers about this story
Vad har hänt?
När hände det?
Varför spelar det roll?
Vilka bolag berörs?
The link opens in a new window and leads to the publisher's own site.
Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.
Get similar news straight to your inbox
The reader's room
Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.
Sign in to submit a comment or question.
Read the article through your role
- Decide whether this affects strategy over 6–12 months or is just noise.
- Discuss with leadership: do we own the right question or does ownership need to move?
- Ask: what risk are we taking by NOT acting on this this quarter?
Generated angle — not editorial analysis of "Open AI Models: New Wave Closes Gap with Closed Systems"