Frontier models cluster at 53 on the Intelligence Index as API margins fall
The author notes that three flagship models all score 53 on the Artificial Analysis Intelligence Index, arguing that frontier capability is converging and API price competition will push inference margins toward zero.
Gemini 4 Argon hitting 53 on the Artificial Analysis Intelligence Index at one-third the cost of Claude Opus 5.5 sounds like a major pricing breakthrough until you check the rest of the chart. Opus 5.5 holds the high mark, but right behind it sits an unbroken knot: • Claude Fable 5.1: 53 • GPT-6 Astra: 53 • Gemini 4 Argon: 53 Comparing Gemini 4 Argon against Opus 5.5 makes Google look aggressive on inference pricing. But set against GPT-6.1 Sol, Argon is still more than twice the price for similar performance. When frontier models cluster at the exact same benchmark score, the illusion of a technical moat evaporates. Measuring affordability against the priciest outlier on the board hides what is actually playing out: frontier intelligence is standardizing, and API margins are getting competed straight down to zero.