RouterPlus News
Model launches, price changes and deprecations, with sources and live catalog prices.
ChatGPT Pro and Business now include up to $500 in API credits
Pro gets $100 to $500 a month, Business $20 or $100 per seat. Unused credits expire each cycle.
OpenAI now gives eligible ChatGPT Pro and Business subscribers free monthly API credits, up to $500, in one linked API organization. The credits expire each billing cycle and do not cover Codex usage through API keys.
Akamai Inference launches in beta with 12 open-weight models
Serverless and dedicated open-model serving on Akamai Cloud. The launch post lists no prices.
Akamai launched Akamai Inference, a managed service for open-weight models behind OpenAI-compatible APIs, in beta on October 27, 2026. The catalog starts with 12 open-weight models, and the announcement gives no per-token prices.
Gemini API now stores requests by default for new paid projects
New paid-tier projects default to store=true. Existing projects keep their setting.
On October 26, 2026, Google made request storage the default for new paid Gemini API projects on the Generate Content API. Teams that create new projects should check the setting before sending production data.
Gemini 4 Carbon announced at $2 input, $10 output
Google's new frontier model keeps Argon's introductory price. Developer access has no date yet.
Google DeepMind announced Gemini 4 Carbon on October 28, 2026, at $2 input and $10 output per million tokens. Access starts with a subset of cyber defenders, with developers to follow.
DeepSeek open-sources DeepInfer, its production inference engine
Apache 2.0, one model, two kinds of hardware, and no KV cache or load balancing.
DeepSeek published DeepInfer, the inference framework behind its production API, on October 20, 2026. It supports only DeepSeek-V4.1-Flash and leaves out the parts that make a full service.