Gemini 3.1 Pro Deep Research warns an oversized prompt may take up to 60 minutes
gaganghotra_ · x · 2026-11-01
Developer gaganghotra shared that a single heavily-loaded prompt to Gemini 3.1 Pro (High) Deep Research triggered an unusual warning: "Putting your report together may take up to 60 minutes, and I'll let you know as soon as it's ready. In the meantime, feel free to leave this chat."
A screenshot shows the Pro High mode with the Start research button, the model openly flagging that the task exceeds its typical runtime — a data point on how long Deep Research actually runs on very heavy requests.
More from Models
- vLLM adopts Strata-style trick: 51B-param Engram lookup table runs without residency in VRAM — solyarisoftware · 2026-11-01
- Leak: GPT-6.2 release imminent, API screenshot shows internal codename astra — kimmonismus · 2026-11-01
- GPT-6.2 Astra Spotted in Codex Usage Analytics as OpenAI Quietly A/B Tests New Model — kimmonismus · 2026-11-01
- 6.1 Sol Good Enough as Daily Driver, Blogger Says 2027 Belongs to Physical AI — haider1 · 2026-11-01
- Dev rants: ChatGPT still won't let you create persistent skill files it can access — JeremyNguyenPhD · 2026-11-01
- Astra 6.2 oddly substitutes Russian letter 'п' for 'p' mid-sentence — teortaxesTex · 2026-11-01