llm Releases Command-Line Client Version 0.31.1
Simon Willison released llm 0.31.1 on July 9, 2026, fixing an OpenAI Chat Completions tool-call bug that could trigger JSON errors when empty arguments were returned by some providers. The update is small, but it matters for practitioners who use CLI-first LLM tooling in scripts, CI jobs, notebooks or local experiments. A narrow provider-compatibility fix can prevent brittle automation failures, especially when teams wrap model calls in shell pipelines. Fedora, Arch and GitHub metadata also show llm remains packaged and discoverable across common developer environments, lowering installation friction for Linux-based workflows.
The practical importance of this release is reliability in automation: small CLI fixes can prevent model-provider edge cases from breaking scripts, demos and CI jobs.
What happened
Simon Willison's release note for llm 0.31.1 says the update fixes a bug where OpenAI Chat Completions endpoints could return a JSON error when a tool call had empty arguments from some providers. The release note says the issue appeared during testing with llm-meta-ai.
Technical context
llm is a command-line tool and Python library for interacting with remote and local large language models. The GitHub project describes support for OpenAI, Anthropic, Google, Meta and other providers, while Fedora and Arch package records show distribution packaging around the 0.31 line.
For practitioners
CLI-first model clients are useful when teams need lightweight glue for data pipelines, evaluations, shell scripts or reproducible experiments without writing a full application. Provider-specific edge cases, especially around tool calls, should be covered by smoke tests before teams rely on the CLI in scheduled jobs.
What to watch
Watch the upstream changelog for provider adapters, Responses API support and plugin compatibility. Package-manager versions may lag upstream releases, so production images should pin the exact version used in tests.
Key Points
- 1The llm 0.31.1 release fixes a narrow OpenAI tool-call bug affecting some provider response payloads.
- 2CLI model clients remain useful glue for scripts, CI jobs and reproducible experiments that do not need full SDK integration.
- 3Linux package records improve install visibility, but upstream release notes remain the authoritative source for behavior changes.
Scoring Rationale
This is a useful tooling maintenance release for practitioners who script LLM calls or rely on CLI automation. The impact is moderate because the change is narrow and does not introduce a major platform or model capability.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
