Tencent Cloud now lists the production Hy4 model as hy4 in TokenHub with a one-million-token context window, 960,000-token input limit, 64,000-token output limit, image understanding, structured output, function calling and caching. The earlier Apache-2.0 Hy4 preview weights remain public, but Tencent has not yet posted a distinct full Hy4 checkpoint and licence.
Why it matters: Teams can now test a documented hosted endpoint with concrete context and tool capabilities, while the difference between the production service and public preview checkpoint remains material for portability and auditability.
Stripe and OpenRouter confirmed their acquisition agreement on 19 August. OpenRouter says its name, mission, product, roadmap, provider choice, and user-driven routing will continue unchanged; it reports more than ten trillion daily tokens across over four hundred models and ten million developers and companies.
Why it matters: First-party confirmation removes deal-existence uncertainty but not concentration risk. A payments company is set to own a major model-access, billing, observability, and routing layer while customers still need portable evidence and tested alternatives.
OpenRouter launched live leaderboards and an open harness comparing model, search engine, method, and one-, five-, or twenty-five-search budgets across BrowseComp, DeepSearchQA, WideSearch, and Humanity's Last Exam. Results expose per-question score, cost, and speed for production endpoints.
Why it matters: The release treats search depth and orchestration as measurable system choices instead of attributing retrieval quality to a model or engine alone.