Ad
Skip to content

Google launches Imagen 3 and two million token context windows in its Vertex AI cloud

Image description
Midjourney prompted by THE DECODER

Google introduces new products for its Vertex AI Cloud.

Gemini 1.5 Flash combines low latency, competitive pricing, and a 1 million token context window. Google says this makes it an excellent choice for various use cases such as retail chatbots, document processing, and research agents.

Gemini 1.5 Pro offers a context window of up to two million tokens, making it suitable for multimodal use cases like analyzing a lot of code, audio files, or videos. Users should be aware of the "lost in the middle" problem that the Gemini models are also likely to suffer from.

Google says Imagen 3, its latest image generation model, generates images over 40 percent faster than its predecessor and follows prompts better. I was able to briefly test the model in a preview version and didn't get the impression that it comes close to current industry-leading generators like Ideogram or Midjourney in terms of quality and prompt accuracy.

Image: Google

Access to Imagen 3 in the Vertex Cloud can be requested here. All generated images are automatically labeled with Deepmind's SynthID.

Google is also expanding the selection of third-party and open-source models in Vertex AI, reducing costs through context caching, improving AI data grounding, and releasing Gemma 2 as a powerful open-source model. Read all the news here.

AI News Without the Hype – Curated by Humans

Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.

Read on for the full picture.
Subscribe for hype-free coverage.

  • Access to all THE DECODER articles.
  • Read without distractions – no Google ads.
  • Access to comments and community discussions.
  • Weekly AI newsletter.
  • 6 times a year: “AI Radar” – deep dives on key AI topics.
  • Up to 25 % off on KI Pro online events.
  • Access to our full ten-year archive.
  • Get the latest AI news from The Decoder.
Subscribe to The Decoder