Google's Gemini 4 Carbon retakes frontier lead two weeks after Argon on new private-heavy benchmark

Google DeepMind retook the frontier lead on October 28 with Gemini 4 Carbon scoring 59 on the new Artificial Analysis Intelligence Index v5, nine points ahead of Argon, just two weeks after Argon's launch. The new index, now 60% private test sets, widened the closed-versus-open gap to 12 points, while Anthropic's October 26 S-1 filing revealed the scale of the race with $16.3 billion in first-half revenue and roughly $780-810 billion in long-term compute commitments. We are confident on Carbon, index and filing facts from lab posts and the S-1, pending independent verification of cyber and benchmark gains.

Outlook Competition likely to stay intense on benchmarks and cyber claims in the next two weeks as labs chase independent verification of Carbon and Index v5 results; pricing responses are unlikely before Carbon's introductory period ends after November 30. Anthropic's compute commitments make a pullback in frontier spending remote, but sustained revenue growth at ~900% retention is unlikely to persist without broader enterprise adoption. Independent tests on Terminal-Bench, FrontierCode and LMArena will likely determine whether Carbon's lead holds beyond lab-reported scores.

Google DeepMind on October 27 added a new front to the frontier race with Genie 4, its first real-time world model with native audio and two-minute scene consistency, expanding competition beyond language benchmarks. OpenAI's October 26 GPT-6.1 Terra price halving to $1/$5 per million tokens still frames the price war, with a new 78.1% EBR-bench record supporting its efficiency claims. No new pricing responses or AGI timeline shifts appeared since October 27; we are confident on Genie 4 and Terra facts from lab posts, pending independent verification of world-model and benchmark claims.

Outlook Genie 4 is likely to shift near-term attention to world-model evaluation (consistency, audio fidelity, steerability) alongside language benchmarks, but adoption will hinge on independent testing after the trusted-tester preview. The Terra price cut makes a further pricing response from Anthropic or Google by November 30 roughly even, while verification of Terra and Kimi on LMArena/APEX and Terminal-Bench is likely within days as API use widens. The ETH refutation claim is unlikely to be settled quickly and will likely remain in community review.

OpenAI's GPT-6.1 Terra on October 26 halved API prices to $1/$5 per million tokens while gaining on larger models on Terminal-Bench and Agents' Last Exam, intensifying a price war that already includes Kimi, Mercury 3 and Gemini 4 Argon. AGI timelines were pulled forward to 2027 by Anthropic's Jack Clark and to 2-3 years by DeepMind's Demis Hassabis. We are confident on pricing and benchmark facts from lab posts; independent verification of Kimi and Terra claims is still pending.

Outlook Price competition is likely to intensify further through November as Gemini 4 Argon's introductory $2/$10 rate is set to double after November 30 and OpenAI's halved Terra pricing pressures rivals to respond. Independent verification of Terra and Kimi K3.1 Pro on LMArena, APEX and wider API use is likely within weeks and will test lab claims. AGI timeline pull-forwards are unlikely to change governance or safety testing pace in the near term without new model releases or incidents.