llm-echo 0.5 has been released on PyPI. The package is a debug plugin for Simon Willison's LLM command-line tool: it provides an echo model that returns prompt details as JSON without calling an API or running a local model.
Version 0.5 documents a -o thinking 1 option. When enabled, the echo model emits two fixed reasoning chunks before its usual JSON text output. The package says this can be used to test software that streams reasoning tokens or persists ReasoningPart objects. It also supports simulated tool calls, raw string responses, and an echo-needs-key model for testing API-key resolution. The release requires Python 3.11 or later and is Apache-2.0 licensed.
PyPI lists the release date as October 18, 2026. The supplied package page does not establish that the canned reasoning output behaves like reasoning from a production model, nor does it provide performance, reliability, or compatibility results for applications using it.