llm-gemini 0.32a0

2026-05-19 · Source: Simon Willison's Weblog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

The "llm-gemini 0.32a0" plugin, released on May 19th, 2026, provides access to Google's Gemini family of large language models. This specific alpha release introduces compatibility with "llm>=0.32a0", an alpha version of the core LLM tool. A significant new feature in this update is the capability to stream reasoning tokens, enhancing the interactive experience and potentially improving the performance or responsiveness of applications built using the Gemini models via this plugin. This allows developers to observe the model's thought process as it generates responses, which can be valuable for debugging and understanding model behavior. The plugin facilitates integration of Gemini models into workflows that utilize the "llm" command-line tool and Python library.

Key takeaway

For AI Engineers integrating Google's Gemini models, upgrading to "llm-gemini 0.32a0" is crucial to access the new streaming reasoning tokens feature. This enhancement lets you observe the model's intermediate thought processes, significantly aiding in debugging complex prompts and understanding model behavior. Consider updating your "llm" core library to version "0.32a0" or higher to utilize this capability for more transparent and controllable LLM interactions.

Key insights

The "llm-gemini" plugin now supports streaming reasoning tokens from Google's Gemini models via "llm>=0.32a0".

In practice

• Integrate Gemini models with "llm" tool
• Utilize streaming reasoning tokens

Topics

• LLM plugins
• Google Gemini
• Streaming tokens
• llm command-line tool
• Model debugging
• Release notes

Code references

• simonw/llm-gemini

Best for: AI Engineer, Machine Learning Engineer, NLP Engineer

Related on AIssential

• llm-gemini 0.32 — The llm-gemini 0.32 plugin now supports Google's gemini-3.5-flash model for faster LLM access.
• llm 0.32a2 — OpenAI's GPT-5 class models now use `/v1/responses` for interleaved reasoning, enhancing tool call visibility.
• 🗞️ Google releases Gemini 3 Deep Think, tops ARC-AGI 2 Benchmark With 84.6% — Advanced AI models are achieving human-level performance in complex reasoning and code generation through enhanced architectural and operational methods.
• Evaluation of LLMs for Mathematical Formalization in Lean — LLMs like Gemini 3.1 Pro and Claude Opus 4.7 show high effectiveness in generating formal Lean 4 proofs, with some models offering cost efficiency.
• GPT 5.5 vs Opus 4.8 vs Gemini 3.5 - Which Model Should You Use? — Optimal AI model selection requires matching specific model strengths to task requirements, as no single model excels universally.
• 😺 Google’s sharpest brain yet? 🧠 — Google's Gemini 3.1 Pro now leads in overall intelligence and factual accuracy, offering a cost-effective, advanced AI model.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.