Gemini 4 Argon hits 1M-token outputs, but only in background mode and 64k chunks
daniel_mac8 · x · 2026-10-14
- Gemini 4 Argon's most interesting feature is support for up to 1M token outputs — about 8x the 128k limit of frontier GPT and Claude models, roughly 10 copies of "Harry Potter: The Sorcerer's Stone".
- The catch: it only works in "background mode" and generates outputs in 64k chunks.
- The author also notes Muse Wonder Full supports 944k outputs and wonders if anyone has pushed the limits yet.
Related event: Gemini 4 Argon reportedly supports up to 1M-token output(3 posts)→
More from Models
- Early users wowed by Gemini 4 Argon's 3D design quality — prajdabre · 2026-10-14
- EmbeddingGemma2 on CPU matches OpenAI's embedding-3-large in 96-config retrieval benchmark — tomaarsen · 2026-10-14
- Physical Intelligence launches Pi API beta for fine-tuning π robotic foundation models — chelseabfinn · 2026-10-14
- Researcher says he had Argon attempt the Indus Valley script, with limited progress — adityakusupati · 2026-10-14
- Leaked Gemini 4 Argon benchmark shows trends matching Astra — scaling01 · 2026-10-14
- Gemini 4 Argon beats GPT-6-Astra and Opus 5.5 on expert medical benchmark — toshv · 2026-10-14