Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5
Carbon costs $7.76 per benchmark task at standard pricing, about 23% less than Claude Opus 5.5, but access remains limited to selected users.
At high reasoning, it scores 59 on Intelligence Index v5 and leads four of its nine evaluations. Its coding agent score also reaches 59, three points behind Claude Opus 5.5, while knowledge work improves substantially over Gemini 4 Argon but trails several rivals. On AA-Omniscience, accuracy rises to 58%, though its hallucination rate increases from Argon’s 4% to 11%. Google is offering a 50% launch discount with no confirmed end date; the article’s cost comparisons use standard prices.
HN · Frontpage AI's note
At high reasoning, it scores 59 on Intelligence Index v5 and leads four of its nine evaluations. Its coding agent score also reaches 59, three points behind Claude Opus 5.5, while knowledge work improves substantially over Gemini 4 Argon but trails several rivals. On AA-Omniscience, accuracy rises to 58%, though its hallucination rate increases from Argon’s 4% to 11%. Google is offering a 50% launch discount with no confirmed end date; the article’s cost comparisons use standard prices.
HN · Frontpage AI's note
score 8