611 Upvotes
Qwen3.5-OmniA native omni model for voice, video, and tools
Voxtral TTS by Mistral AIMultilingual TTS model with realistic and expressive speech
MTIA 300Meta's 3rd-gen custom AI chips for GenAI inference
Qwen-Image-2512SOTA open-source T2I model with even greater realism
Qwen-Image-LayeredTurn flat images into multi-layer editable assets
SAM AudioSegment any sound with text, visual, or time prompts
Devstral 2SOTA open-source agentic coding models and CLI agent
Seedream 4.5High-fidelity multi-image editing & dense text rendering
Manus Browser OperatorAny browser can now become an AI browser
Kimi K2 ThinkingThe 1T Parameters Open-Source Thinking Model - SOTA on HLE
Pomelli by Google LabsYour copilot for on-brand content at any scale
Meta Ray-Ban DisplayA breakthrough category of AI glasses
Google Finance BetaDive into the world of finance with AI-powered insights
Grok 2.5 (OSS Ver.)2024 best model from xAI, now open source.
OpalDescribe, create, and share your AI mini-apps
Qwen3-CoderA powerful open model for agentic coding tasks
ComfyUI-CopilotBuild ComfyUI workflows with natural language
VoxtralFrontier open source speech understanding models
Gemini Robotics On-DeviceGoogle's best robotics AI for the edge
11.ai by ElevenLabsThe voice-first AI assistant that takes action