llm 0.31.1

The latest version of LLM, a command-line utility and Python library for working with large language models, has been released with an important bug fix targeting OpenAI Chat Completion endpoints. Version 0.31.1 resolves an issue where tool calls containing empty arguments could trigger JSON parsing errors from certain providers, ensuring more reliable integration across different LLM services.

LLM 0.31.1 specifically addresses bug #1521, which surfaced during testing of llm-meta-ai compatibility. The issue occurred when language models made function calls with empty argument parameters, causing downstream JSON parsing failures in some OpenAI Chat Completion implementations. This fix ensures that tool calls are properly formatted regardless of whether arguments are present or empty, improving the robustness of LLM's interface with external AI providers.

• Improved API Reliability: The fix prevents JSON errors that could disrupt automated workflows relying on tool-enabled LLM integrations
• Better Provider Compatibility: Addressing edge cases like empty arguments enhances cross-provider consistency and reduces implementation-specific errors
• Enhanced Testing Coverage: The discovery during llm-meta-ai testing demonstrates the value of comprehensive integration testing across multiple LLM platforms
• Developer Experience: Developers using LLM for function calling now face fewer unpredictable failures, reducing debugging time and improving development velocity
• Tool Use Standardization: The fix contributes to more standardized handling of function calls across different LLM implementations

While seemingly technical, this release represents an important maintenance milestone for the LLM ecosystem. Tool calling—the ability for language models to invoke external functions—represents a critical capability for building practical AI applications. By eliminating a subtle but consequential bug, version 0.31.1 strengthens the foundation upon which developers build AI-powered applications. For teams leveraging LLM as their primary interface to language models, particularly those integrating Meta AI capabilities, this update removes a potential source of production issues. The fix underscores the ongoing importance of rigorous testing in AI tooling and the cumulative benefits of regular maintenance releases.