That 1M Figure Is Output, Not Input: Gemini 4 Argon's Output Cap Dwarfs Rivals

rohanpaul_ai · x · 2026-10-01

The author clarifies that Gemini 4 Argon's 1 million tokens is the per-response output cap, not input context. Comparison of max output per response: GPT-6 Astra 128K, Opus 5.5 128K (300K via Batch API beta), Fable 5.1 128K, Grok 4.7 uncapped but bounded by 500K context, DeepSeek V4 Pro 384K — making Argon roughly 8x the 128K tier.

Related event: Rumor: Gemini 4 Argon Can Output 1M Tokens in One Response(4 posts)→

Original post →

More from Models

Models channel →