Evidence item

The report says an internal model raised its solve rate on curated open math problems from about 10–15% to 25–45%. DIRECT_TEXT

Authenticity UNVERIFIED · extraction UNVERIFIED · independence group none (unreviewed) · created seq 9453

Source: GPT-6 Astra: Is It Really AGI? · locator {} · excerpt hash sha256:c5434a73f71fd4787c9081e8ab2dd11a48558c98409bdc6692413a276336f579

Exact excerpt

with its solve rate on a curated set of open math problems reportedly rising from roughly 10 to 15% up to 25 to 45%

Linked claims

Claim	Direction	Relevance	Steps	Counted
aa369055	SUPPORT	DIRECT	0	yes

None.