Hugging Face Inference Providers now send LLM calls via OpenRouter, with retries and :fastest or :cheapest routing. Your own provider keys no longer work.
DeepSeek's 13 November API release adds server-run web search, file parsing, URL reading, TTS, ASR and a safety check to deepseek-flash, with per-use prices.
India joins Australia and Japan for in-country API inference on gpt-6-luna. It needs Zero Data Retention, Standard pay-as-you-go and a 10% residency uplift.
Anthropic's table shows the watermark live on Mythos 5, Opus 4.7 and Sonnet 4.6 since 3 November. Opus 4.6 missed the date; a 16 November wave is reported.
Eight changes from our past week for people building with AI: two security fixes, a new default, a hosted shutdown and new tools, each with a next step.
Mistral's stateless /v1/responses endpoint returns the answer, reasoning and tool calls as separate items, for its hosted chat models including Large 4.
Canva trained Craft from scratch for layouts and lettering in 55 languages. It is live in Canva and in a preview Connect API endpoint that spends AI credits.
OpenAI's updated Usage Policies say services that give third parties direct access to its models need prior review. Training limits also tighten. From 20 Nov.
Anthropic's first API version since 2023 returns a 400 for temperature, top_p or top_k. New major SDKs send it by default. What breaks and how to stay put.
On Friday OpenAI's API retires a batch of older models announced in April, including gpt-image-1 and the o-series minis. Check your model strings today.
Inception says Mercury 3 matches or beats GPT-6 Luna and Claude Haiku 5.5 while running 4 to 8 times faster. API-only, 50% off at launch, benchmarks self-run.
DeepSeek's API guide documents Search and Fetch at 5 requests a second. The Chinese pages price each at 5 yuan per 1,000 successes; the English pages say $0.75.
Some model names are fixed snapshots and some are pointers that move. How Anthropic and OpenAI name their models, how to tell which kind you are calling, and what even a pinned ID does not freeze.