https://clfaigateway.dev/blog Engineering notes from CLF AI Gateway: per-token billing you can audit, caching, and shipping with the OpenAI SDK. en Tue, 18 Aug 2026 00:00:00 GMT https://clfaigateway.dev/blog/which-models-read-images https://clfaigateway.dev/blog/which-models-read-images Qwen 3.8 27B joins the gateway with verified image input. While adding it we probed all seven models for vision — and found three ways a model can fail at seeing. Tue, 18 Aug 2026 00:00:00 GMT product https://clfaigateway.dev/blog/deepseek-v4-1m-context https://clfaigateway.dev/blog/deepseek-v4-1m-context deepseek-v4-flash and deepseek-v4-pro join the gateway: a 1,048,576-token context window, verified function calling, Cloudflare-hosted, 40% off list price. Sun, 16 Aug 2026 00:00:00 GMT product https://clfaigateway.dev/blog/clf-ai-gateway-is-live https://clfaigateway.dev/blog/clf-ai-gateway-is-live The OpenAI-compatible gateway for Kimi and GLM is open to everyone: prepaid credits, per-token billing, and 40% off list prices for the launch period. Wed, 12 Aug 2026 00:00:00 GMT product https://clfaigateway.dev/blog/how-billing-works https://clfaigateway.dev/blog/how-billing-works The exact formula behind every CLF AI Gateway charge: per-token prices, cached-input discounts, zero-completion insurance — with a worked example. Tue, 11 Aug 2026 00:00:00 GMT product https://clfaigateway.dev/blog/ship-with-openai-sdk https://clfaigateway.dev/blog/ship-with-openai-sdk Point the official OpenAI SDKs, LangChain, or the Vercel AI SDK at CLF AI Gateway: swap base_url, keep your code. Streaming and error handling included. Tue, 11 Aug 2026 00:00:00 GMT guides