FlashNEWS BRIEFING

Gemini 4 Cobalt Preview Surfaces in Google AI Studio

1 min read

88Z AI, 22 November — A new model identified as "gemini-cobalt-thinking-preview-001" has appeared in Google AI Studio, featuring multimodal support and up to 131,072 reasoning tokens. Given the preview label and Google's release cadence, a public launch may occur within weeks.

At a glance

  • Gemini 4 Cobalt preview is live on Google AI Studio with support for text, audio-video, and 3D inputs
  • The model offers up to 131,072 reasoning tokens and 4K video CoT reasoning, with a likely release in weeks

Model Appearance and Release Timeline

A new model identifier, "gemini-cobalt-thinking-preview-001," has recently surfaced on the Google AI Studio platform. The inclusion of the "preview" tag suggests that Google is currently testing an early checkpoint of this system. Based on Google's recent release cadence, industry observers speculate that Gemini 4 Cobalt could be publicly launched within the next few weeks.

Core Specifications and Multimodal Capabilities

Leaked technical details indicate that Gemini 4 Cobalt features robust multimodal processing, supporting text, images, audio, video, and 3D content. In terms of reasoning performance, the model supports up to 131,072 reasoning tokens and specifically highlights chain-of-thought (CoT) reasoning for 4K video. While it remains unclear whether Cobalt will belong to the Flash or Pro lineup, its ability to deeply understand high-resolution video has drawn significant attention to its potential visual performance.

Sources & editorial note

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.

  1. @Sancani_Source published

Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.

Related reading

All stories
88Z.ai