Anthropic introduces Video-DeepResearch, a next-generation multimodal deep research agent that extends multimodal agents from static images to continuous video streams. This breakthrough addresses modality bias and parametric knowledge leakage, achieving state-of-the-art results. The model outperforms existing models, including Claude-4.5-Sonnet and GPT-5.

OpenAI GPT-5.6 Sol, Terra, and Luna are now available on Amazon Bedrock, offering developers a range of models for agentic coding, long-horizon reasoning, and high-volume inference workloads. The models provide a balance of performance, cost, and security, and can be accessed through the OpenAI Responses API. With pricing matching OpenAI first-party rates, developers can easily integrate these models into their existing workflows.