Beats every frontier model we tested

Subject: SKY-001

Claim

On CAD and BIM tasks, SKY-001 "beats GPT-6 Astra and every frontier model we tested": 51–125× faster, 150–210× cheaper, and a higher IoU than all frontier models.

Finding

The comparison is the lab's own. On its CAD task the accuracy lead is 79.7% against 79.1%; the speed lead is 3.9 seconds against 199. The 0.6-point lead and the 51× lead were announced in the same breath, which is the correct order if you are selling it. The model is a research preview, and no independent evaluation had been published at the time of this build.

Timeline

19 Oct 2026
Announced: "beats GPT-6 Astra and every frontier model we tested."@spisallyouneed · ♥ 16.4K ↗
22 Oct 2026
"Already outperforms all the frontier LLMs in CAD and BIM."@skyfallai · ♥ 875 ↗
Ongoing
Awaiting a second table.

Primary sources

The posts this entry is based on, embedded as published.

Sam Pasupalak

@spisallyouneed

Introducing SKY-001, the first Engineering World Model. On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested: -> 51-125x faster - > 150-210x cheaper -> Higher IOU than all frontier models SKY-001 lets you dream a physical object before you build it. 18

19 Oct 2026♥ 16.3KOpen on X ↗

Skyfall AI

@skyfallai

We launched the first Engineering World Model, SKY-001, this week and it already outperforms all the frontier LLMs in CAD and BIM in speed, cost and precision. Here’s an example of how our model compares to OpenAI’s Astra, Anthropic’s Fable 5.1 and Google’s Gemini 4. These

22 Oct 2026♥ 1KOpen on X ↗