New function tabs
Web Search for fast information retrieval.
Reasoning for complex problem-solving and analysis.
Deep Research for in-depth analysis and data exploration.
New models
Gemini 2.5 Pro, Google’s most advanced model, strong at reasoning and coding.
Mochi 2.0, Babbily’s retrieval model that combines Llama 3.3 with real-time Google search results.
Boba 2.0, a reasoning model trained on DeepSeek R1 Distill Qwen 14B.
Cohere Command A for more nuanced, context-aware language tasks.
Google Imagen 3.0 for realistic image generation.
Grok 2 image generation for detailed, varied images.
Flux 1.1 Pro for complex and artistic images.
Google Flash multimodal, so you can add images directly in chat.