TrustList insights

Artificial Intelligence News

21 articles, newest first. Last updated 7 Oct 2026.

21 articles

45 open-weight models, 53 closed ones and the licences between them

Eight of the 13 specified models dated October ship downloadable weights, and all eight are Apache 2.0, while only five of the 45 open rows record any price.

7 Oct 2026

Also new

Artificial Intelligence News

Seven decision models in ten days, and not one independent score

27 Sept 2026

Artificial Intelligence News

When the model you build on stops answering

23 Sept 2026

Artificial Intelligence News

The leaderboard carried none of September's thirty model releases

20 Sept 2026

More artificial intelligence news

Artificial Intelligence News

What makes a token expensive, and why it is not the model's size

Our own research, from a board of 101 AI models priced one vendor page at a time: of the 23 that publish a per-token rate, 22 charge a median 4x more for output than input, and only 3 publish a cached rate worth 9-10x.…

19 Sept 2026

CRM Software

Small models are winning the argument and losing the evidence

Our board carries 99 models and 30 benchmarks. Of September's 25 releases, 8 carry a comparable score and 17 do not - and the 17 are mostly the small, specialised ones the trend is celebrating.

18 Sept 2026

CRM Software

Seven of this week's fourteen launches sell to the agent, not the person

Of fourteen business launches on 16 and 17 September, seven are sold to somebody building or running AI agents - and only three of the fourteen still charge per seat.

18 Sept 2026

CRM Software

The best launches this week built the brake before the accelerator

Of eleven business launches that ranked on 8–10 September, four put a human approval step at the centre of the product - and seven were about constraining an agent, not unleashing one.

14 Sept 2026

Business Intelligence

The busiest category on our launch board is the layer between your app and the model

84 launches in 30 days: 14 AI models, and 21 tools to route, govern and meter calls to them.

12 Sept 2026

Artificial Intelligence News

Five releases in four days, and not one of them was really about capability

OpenAI, Sakana AI and DeepSeek all shipped in the same four days. Every release was a pricing decision wearing a model announcement.

12 Sept 2026

Artificial Intelligence News

A resignation, a number, and what "gambling with our lives" actually claims

An Anthropic researcher quit and his colleague put a number on the risk. The number is not the interesting part.

9 Sept 2026

Artificial Intelligence News

What 215 launches in ninety days say about where tooling is going

Eighty-four model releases, 129 product launches, and a centre of gravity that has moved from the model to the runtime around it.

8 Sept 2026

SQL

The semantic layer stopped being a background discipline

A dashboard with a wrong metric definition produces a number a human questions. An agent with a wrong metric definition takes an action.

8 Sept 2026

Data Intelligence

Data intelligence consolidated while everyone was watching the models

Three deals in seven months moved ingestion, transformation, streaming and master data inside platform vendors - and the buyer absorbs the transition.

8 Sept 2026

Artificial Intelligence News

Meta's Contributor tier: tokens up to 21× cheaper, paid for with your data

Muse Spark 1.3 has two prices for the same model. The cheap one sends your prompts and outputs into Meta's training pipeline. Here's how to decide which workloads that's actually fine for.

5 Sept 2026

Cybersecurity

Cyber-capable AI models are now access-gated: Fairwind, Daybreak and Glasswing in one week

Google, OpenAI and Anthropic each shipped a model whose full security capability is available only to vetted organisations. The benchmarked version and the purchasable version are no longer the same product.

5 Sept 2026

Cybersecurity

GPT-6 Astra's 99.9% on ARC-AGI-3 is two numbers, not one

ARC Prize's own report gives 62.7% under the shared harness and 99.9% under OpenAI's adapter - and says neither is proof of AGI. Which one describes what you'll get depends on how you integrate.

5 Sept 2026

Artificial Intelligence News

Four labs, 72 hours: what the September 2026 frontier wave actually changed

Anthropic, Google, Meta and OpenAI all shipped between 1 and 3 September. Read together, three things changed: the premium tier is back, prices now expire, and the most capable variants are gated on purpose.

5 Sept 2026

Artificial Intelligence News

Open weights went frontier-class this summer, so read the licence before you plan on it

Kimi K3, Qwen3.8, GLM-5.3 and Granite 4.2 put serious capability on your own hardware. "Open" is now doing a lot of load-bearing work in the press releases.

31 Aug 2026

Artificial Intelligence News

The August 2026 model wave: what actually changed for software buyers

Every major lab shipped this summer. Three things changed that matter to a budget - and none of them is "the models got smarter".

31 Aug 2026

Purchase Order Management

Open weights are back: Meta reopens the model debate, and buyers get leverage

Meta has pivoted back to open-weight model releases with Muse Glimmer, with Zuckerberg publicly criticising closed rivals. Strip out the rhetoric and there is a real procurement question underneath: when does owning…

10 Aug 2026

TrustList Weekly

The week in software and IT, in one email

The news that matters to buyers, new rankings and our own research. Every Thursday, free, and easy to leave.

Keep exploring

More on TrustList

Everything here links back to the same verified catalogue. Pick your next stop.