Global AI Native Industry Insights – 20260915 – ElevenLabs | Perplexity AI | Cognition | more
ElevenLabs and OpenAI take center stage in today’s digest, showcasing advancements in voice and biological reasoning technologies. A common theme throughout is the expansion of AI capabilities across different modalities, with ElevenLabs adding voice, music, image, and video generation to its MCP server, and OpenAI launching GPT-Rosalind to enhance biological reasoning in API and Codex. Perplexity introduces the Comet Assistant for Windows PCs with NVIDIA RTX GPUs, emphasizing compatibility and performance enhancements. Cognition debuts Fusion, a model harness aimed at cost efficiency for Devin CLI users. Discover more in Today’s Global AI Native Industry Insights.
1. ElevenLabs adds voice, music, image, and video generation to its MCP server
ElevenLabs has expanded its Model Context Protocol (MCP) server to support generation of voice, music, images, and video, in addition to existing speech capabilities. Users can now generate speech, transcripts, dubs, music, sound effects, images, and video directly from within the AI assistant they already use. The update turns the ElevenLabs MCP into a multimodal content generation tool accessible through compatible assistants. This broadens the range of creative and production tasks developers and users can perform without leaving their existing workflow.
Read more: https://elevenlabs.io/mcp
Video Credit: @ElevenLabs on X
2. Perplexity’s Comet Assistant Launches for Windows PCs with NVIDIA RTX GPUs
Perplexity AI announced that its Portable Computer agent tool is now available on Windows PCs equipped with NVIDIA RTX GPUs. The tool allows users to run its harness, agents, and models locally on their own hardware. It supports working with local files and connected apps without sending tasks to the cloud, while still allowing access to frontier cloud models when needed. The release expands local, on-device AI agent capabilities for Windows users.
Read more: https://www.perplexity.ai/hub/products/portable-computer
Video Credit: @perplexity_ai on X
3. Cognition Launches Fusion, a Cost-Efficient Model Harness for Devin CLI
Cognition introduced Fusion, a new harness within Devin CLI that combines separate models for planning and execution. The company says the approach works with its Fable and Astra models and cuts costs by 39% across coding benchmarks compared to using a single frontier model. Users can select a preferred model for planning tasks while pairing it with a more cost-effective model for execution. The release targets developers seeking lower-cost AI coding workflows without sacrificing performance.
Read more: https://cognition.com/blog/local-fusion
Video Credit: NotebookLM
4. OpenAI launches GPT-Rosalind for biological reasoning in API and Codex
OpenAI has introduced GPT-Rosalind, a model designed to bring stronger biological reasoning capabilities to its API and Codex platform. The model helps researchers connect findings across scientific papers and experimental results. It can weigh evidence for a given biological target and work through analyses to help plan future experiments. The release targets researchers working on biological and life sciences problems.
Read more: https://openai.com/index/introducing-gpt-rosalind/
Video Credit: @OpenAIDevs on X
That’s all for today’s Global AI Native Industry Insights. Join us at AI Native Foundation Membership Dashboard for the latest insights on AI Native, or follow our linkedin account at AI Native Foundation and our twitter account at AINativeF.