Skip to main content

Institution profiling / Global Cloud Services

Is Google’s just-released Gemini 1.5 Pro better than ChatGPT-4?

Google’s release of Gemini 1.5, featuring a multi-modal model with 1 million token context processing, disrupts the AI landscape.

Is Google’s just-released Gemini 1.5 Pro better than ChatGPT-4?

Is Google’s just-released Gemini 1.

  • Processing one million tokens at a time gives Gemini 1.5 Pro broader possibilities in business applications;
  • Google just released Gemini 1.0 in December last year, and this sudden update is clearly “prepared.”

Previously, most would point to OpenAI’s GPT-4 Turbo as Earth’s mightiest AI model. However, Google‘s sudden unveiling of the “next-gen AI model” Gemini 1.5 has swiftly shifted perceptions, boasting remarkable productivity enhancements compared to its predecessor, Gemini 1.0, launched just last December.

A significant upgrade of a multi-modal universal model

The highlight of this new version is the debut of Gemini 1.5 Pro, a multi-modal universal model that expands the upper limit of stable context processing to 1 million tokens, ushering in an era where AI models can handle text, code, images, voice, and video, among other rich media types.

This significant upgrade surpasses the context understanding limits of both Gemini 1.0 Pro, released two months prior with a context understanding cap of 32,000 tokens, and OpenAI’s GPT-4 Turbo, which supports only 128,000 tokens. To put it simply, Gemini 1.5 Pro can process over 700,000 words of text, 30,000 lines of code, 11 hours of audio, or 1 hour of video in a single instance, thanks to its “multi-expert model” (MoE) algorithm, which runs only a portion of the overall model to respond to requests faster and more efficiently.

Also read: Google’s Bard chatbot gets the Gemini Pro update globally

Future AI models could handle the entire ‘Game of Thrones’ series in one go

Google CEO Sundar Pichai also revealed that Google’s researchers have successfully tested a context window of 10 million tokens, indicating future AI models could handle the entire “Game of Thrones” series in one go (totalling 1.73 million words across five published books). Currently, the testing version of Gemini 1.5 Pro offered to developers has a token limit of only 128,000, but DeepMind CEO Demis Hassabis bluntly stated that new paid tiers unlocking the 1 million token version will be introduced later. As a comparison, the current subscription fee for using the Google Gemini 1.0 Pro model is $19.99 per month.

Member Briefing

Read the full profile briefing

Members read the full briefing with the source notes behind every finding.

Only for Strategic Circle

Strategic Circle

Join to read every profile briefings in full, with the source notes behind them.

Join Strategic Circle

Only for Leadership Alliance

Leadership Alliance

For IP-asset owners and leadership teams: alliance briefings, private reports and direct analyst support.

Join Leadership Alliance
Snapshot
Confidence
Confidence score guide
Limited confidence (76%)

Several public sources

BackAll Companies Only for Strategic Circle

Recommended reading

Only one of Google's three per-prompt climate numbers came off a meter
Global Cloud ServicesOnly one of Google's three per-prompt climate numbers came off a meter18 Nov 2026
Same Metric Name, Two Different Denominators
Global Cloud ServicesSame Metric Name, Two Different Denominators18 Nov 2026
Microsoft's $207 Million OpenAI Gain Has No Line to Sit In
Global Cloud ServicesMicrosoft's $207 Million OpenAI Gain Has No Line to Sit In17 Nov 2026
Every Ciena Terabit Arrives With a Distance Attached
Global Cloud ServicesEvery Ciena Terabit Arrives With a Distance Attached13 Nov 2026
Akamai Rents Out $109 Million of Cloud a Quarter and Spends $464 Million Building It
Global Cloud ServicesAkamai Rents Out $109 Million of Cloud a Quarter and Spends $464 Million Building It13 Nov 2026