Release Event And Version Boundary

Issue 90 Edition 2026-03-31 3 min read

Not accepted General

Sources: 1 • Confidence: Medium • Updated: 2026-04-12 10:21

Key takeaways

llm-echo version 0.4 has been released.
In llm-echo 0.4, prompt responses include populated input_tokens and output_tokens fields.

In llm-echo 0.4, prompt responses include populated input_tokens and output_tokens fields.

What other changes (breaking or non-breaking) are included in llm-echo 0.4 besides token-field population?
Across which providers/models are input_tokens and output_tokens populated, and are the values exact, estimated, or occasionally missing?
What is the schema/contract for token fields (data types, naming, units, and backward compatibility with prior response formats)?
Is there any direct decision-readthrough (operator, product, or investor) beyond “test and consider upgrading” for token telemetry needs?

Token fields in responses could enable better usage analytics and cost attribution for teams using llm-echo, potentially increasing product stickiness for telemetry driven workflows.
Version 0.4 release creates an upgrade boundary that could prompt evaluation of compatibility and operational changes, affecting adoption pace depending on undocumented breaking changes.

Changelog shows limited, non breaking changes beyond token field population and clear backward compatibility guidance.
Documentation specifies token field schema, units, and reliability and confirms broad provider and model coverage with exact or consistently estimated counts.
User reports or release notes highlight downstream tooling adoption for cost tracking, analytics, or rate limit management enabled by the new token metadata.

Token fields are frequently missing or inconsistent across providers or models, limiting usefulness for telemetry and cost tracking.
Breaking changes or unclear migration requirements in 0.4 increase upgrade friction and slow adoption.
Schema for token fields is unstable or changes across patches, undermining backward compatibility and downstream integrations.