StoryFirst seen
Its switch to closed benchmarks shows a wider gap than the company had measured before.
Everything it touches
1 names 1 voices 1 source posts
The story’s neighborhood
Explore the graph, or choose a connection below.
Follow the evidence
This brief comes from Tech Twitter’s edited headlines. The posts below are the sources grouped with the story; statements in them belong to their authors.
Artificial Analysis finally moves to closed benchmarks (goodbye GDPval-AA), and finds that open weights models are falling further behind than they had previously measured. Take with a grain of salt, but matches the general vibes that we haven't seen a Fable-class open model.Read original on X
Artificial Analysis says open-weight models trail frontier systems by more
Artificial Analysis finds open-weight models falling further behind frontier systems
Artificial Analysis says open-weight models fall further behind frontier systems
Artificial Analysis finds open-weight models trail frontier systems further
Artificial Analysis finds open-weight models trail frontier systems by more
Artificial Analysis says open-weight models trail frontier systems by more
Artificial Analysis finds open-weight models falling further behind frontier systems
Shared connections
If confirmed, the claim would challenge a basic assumption about how hard a common class of logic problems is to solve.
Both labs' models independently reached a faster solving bound for a hard logic problem, though this does not settle P versus NP.
The filing also shows an 85% gross margin on existing models while the company spends heavily on rented computing power.
The filing gives investors a detailed look at the AI lab's finances, revenue, and spending obligations.