Model Intelligence

Gemini 1.5 Pro

Name: Gemini 1.5 Pro
Author: Google DeepMind

Frontier Multimodalby Google DeepMindReleased February 2024

Gemini 1.5 Pro introduced a 1 million token context window — a structural step change in capability for long-document, multi-document, and long-video reasoning tasks. Built on Google's Mixture of Experts architecture, it maintains competitive performance while dramatically extending usable context beyond any prior frontier model.

Context Window

1,000,000 tokens (2M in preview)

License

Proprietary (Google AI API + Gemini.app)

Capabilities

1M token context window (industry leading)
Multimodal: text, image, video, audio, code
In-context learning from very long documents
Video understanding (1 hour+ videos)
Multi-document analysis and synthesis
Code generation with workspace understanding

Pricing Summary

Tiered by context length. Up to 128K: $3.50/1M input. Above 128K: $7.00/1M input. Output: $10.50/1M tokens.

Benchmarks

MMLU: 85.9% | HumanEval: 84.1% | MATH: 67.7% | MRCR (1M): 99.7%

Competitive Positioning

Google's primary competitive advantage is context length. Workloads requiring analysis of entire codebases, legal documents, or long video content are Gemini's clearest moat. Deployed natively in Google Workspace giving unmatched distribution to enterprise knowledge workers.

Related Models

Gemini 1.0 UltraGemini 1.5 FlashGemini NanoPaLM 2