OpenAI ships new frontier model with improved coding and vision
OpenAI's latest release focuses on better coding, longer context and enterprise reliability. Early benchmarks show a meaningful jump on SWE-bench and MMLU-Pro.
Newsroom
The most important AI stories — model releases, product launches, funding and research — updated weekly.
OpenAI's latest release focuses on better coding, longer context and enterprise reliability. Early benchmarks show a meaningful jump on SWE-bench and MMLU-Pro.
Claude's new release adds richer tool use, faster streaming and improved coding — closing the gap with GPT on agentic tasks.
Gemini 2 Pro is now the default across Google Workspace, with improved reasoning and Veo 3 integration for enterprise customers.
Perplexity has raised a mega-round led by top-tier investors, doubling its valuation. Deep Research and Spaces are cited as key growth drivers.
Mistral's new open-weight release outperforms comparable-size models on reasoning and multilingual benchmarks.
Meta previews an agent stack built on Llama, targeting enterprise workflows and on-device inference.
The European Commission has published enforcement guidance for foundation model providers under the AI Act.
AI-first IDE Cursor has crossed a major ARR milestone, becoming one of the fastest-growing developer tools ever.
ElevenLabs' new Realtime API brings latency competitive with OpenAI's voice offering while retaining voice cloning quality.
Veo 3's latest update improves synchronized dialogue and ambient sound generation for creators.