Follow the latest coverage, related explainers and connected technology stories.
GitHub explains that reducing tokens in each invocation does not necessarily lower the cost of the entire task, as it may prompt the model to rerun commands or retrieve deleted outputs. The company presents four improvements in Copilot: compressing repetitive outputs, removing unhelpful formatting, shortening instructions, and delivering background-task results directly.
DeepSeek has launched the official version of the V4-Pro model alongside a preview of the open-source DeepSeek Harness, expanding its presence from the model layer to the infrastructure that manages coding agents’ tools. The move comes with support for the Responses API and Codex, but also with significant increases in API pricing depending on peak periods.