Three AI labs account for 37% of every model launch on TrustList's board
EditorialBy TrustList Editorial
OpenAI, Google and Anthropic are shipping far more often than the rest of the field. What launch frequency actually tells a buyer — and what it doesn't.
About Three AI labs account for 37% of every model launch on TrustList's board
Group TrustList's 76 AI-model launches by the lab behind them and the board stops looking like a level field. Three labs account for 37% of every model launch recorded — and the next tier down is a tight cluster of labs shipping at a noticeably slower, steadier pace.
The real count, by lab
| Lab | Launches recorded |
|---|---|
| OpenAI | 12 |
| 9 | |
| Anthropic | 7 |
| Alibaba (Qwen) | 5 |
| DeepSeek | 4 |
| Meta | 4 |
| Z.ai (Zhipu AI) | 3 |
| IBM | 3 |
| xAI | 3 |
| Mistral AI | 3 |
| Cohere | 2 |
Past that, twenty-one further labs — Moonshot AI, Stability AI, Adobe, Midjourney, NVIDIA, Amazon, Microsoft and others — each show exactly one launch in the window. That long tail is not a sign those labs are inactive; it reflects that most labs ship a flagship generation occasionally, while the top three are effectively always shipping something.
What "12 launches" from one lab actually means
It does not mean twelve unrelated model families. It means a cadence of variants, updates and specialised releases layered on top of a smaller number of true generational shifts — the same pattern visible across OpenAI, Google and Anthropic's entries on this board. For a buyer, the practical read is: a lab appearing repeatedly on the launch board is iterating publicly and often, which is useful to know when deciding how much to invest in integrating against any one specific model version versus the lab's API surface as a whole.
The three-lab concentration, and its limit as a signal
37% of launches from three labs is a real concentration, but it measures announcement frequency, not capability, safety or value for a given task — a lab with one launch in this window can still be the right choice for a specific job. What the concentration usefully tells a buyer is where to expect the fastest model churn: if you're building against OpenAI, Google or Anthropic's APIs, budget for more frequent model-version decisions than you would with a lab shipping once a quarter.
Launch counts computed from TrustList's live launch board
(/launches/board, window=all, type=ai_model) on 31 August 2026: 76 total
AI-model launches, attributed to their lab by model name.
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- AI ModelsBenchmark scores and community ratings for every major model.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.