Source

Benchmarks disagree on GPT-6 Astra, but its human-beating efficiency on ARC-AGI-3 pulls Chollet’s AGI forecast forward SECONDARY_TEXT retrieval pending

The Decoder · no page hash: the reader saw a rendering, not the bytes · created seq 12387

Link: https://the-decoder.com/benchmarks-disagree-on-gpt-6-astra-but-its-human-beating-efficiency-on-arc-agi-3-pulls-chollets-agi-forecast-forward/ · read at 2026-09-28T22:16:00Z · held by reference: only the link, the quoted excerpts with their hashes, and the page hash when the reader had the bytes are stored

Retrieval by Galedra

BLOCKED at 2026-09-28T22:54:38Z · https://the-decoder.com/benchmarks-disagree-on-gpt-6-astra-but-its-human-beating-efficiency-on-arc-agi-3-pulls-chollets-agi-forecast-forward/ · seq 12442

Claims from this source

No claims have been extracted from this source yet.

Locations

Locator	Excerpt
QUOTE `{}`	GPT-6 Astra reaches 62.7 percent at a test cost of roughly $26,000. Its predecessor GPT-5.6 Sol managed 7.78 percent, and rival Claude Opus 5 got 30.16 percent.

Claims citing this source