Keimodel
Log inCompare
Google

Google

Multimodal AI with million-token context windows

Google's Gemini family represents the company's most capable AI models, built natively multimodal from the ground up. Gemini 2.5 Pro and Flash lead on long-context tasks with their 1M-token context windows, enabling analysis of entire codebases, lengthy documents, and extensive conversations. Google's TPU infrastructure provides competitive pricing at frontier capability levels.

Founded 1998·5 models·Visit Google

5 Models

Gemini 3.5 Pro

Google's most capable next-gen model

5 cr
Gemini 3.5 Pro is Google's frontier model in the 3.5 generation, combining the best of multimodal reasoning, coding, and long-context analysis with a 1 million-token context window. It targets the hardest research, engineering, and analytical tasks.
1M ctxVisionTool use

Gemini 3.5 Flash

Google's next-gen fast multimodal model

3 cr
Gemini 3.5 Flash is Google's latest generation fast model, combining improved reasoning and multimodal understanding with low latency and a 1 million-token context window. It is the recommended default for most everyday tasks.
1M ctxVisionTool use

Gemini 2.5 Pro

Google's most capable model with 1M context

5 cr
Gemini 2.5 Pro is Google's frontier model, offering native multimodal understanding, a 1 million-token context window, and top performance on coding, mathematics, and scientific reasoning benchmarks.
1M ctxVisionTool use

Gemini 2.5 Flash

Fast multimodal model with 1M context

3 cr
Gemini 2.5 Flash delivers near-Pro performance at a fraction of the cost. Its 1M-token context window and low latency make it the go-to model for real-time applications, summarisation, and document processing at scale.
1M ctxVisionTool use

Gemini 2.5 Flash Lite

Ultra-efficient with massive context

2 cr
Gemini 2.5 Flash Lite is Google's most cost-effective model, retaining a 1M-token context window while delivering rapid responses at the lowest price point in the Gemini family.
1M ctxVisionTool use

Other providers

OpenAIOpenAIAnthropicAnthropicMetaMetaMistralMistral