Source URL: https://simonwillison.net/2025/Sep/21/llm-openrouter/
Source: Simon Willison’s Weblog
Title: llm-openrouter 0.5
Feedly Summary: llm-openrouter 0.5
New release of my LLM plugin for accessing models made available via OpenRouter. The release notes in full:
Support for tool calling. Thanks, James Sanford. #43
Support for reasoning options, for example llm -m openrouter/openai/gpt-5 ‘prove dogs exist’ -o reasoning_effort medium. #45
Tool calling is a really big deal, as it means you can now use the plugin to try out tools (and build agents, if you like) against any of the 179 tool-enabled models on that platform:
llm install llm-openrouter
llm keys set openrouter
# Paste key here
llm models –tools | grep ‘OpenRouter:’ | wc -l
# Outputs 179
Quite a few of the models hosted on OpenRouter can be accessed for free. Here’s a tool-usage example using the llm-tools-datasette plugin against the new Grok 4 Fast model:
llm install llm-tools-datasette
llm -m openrouter/x-ai/grok-4-fast:free -T ‘Datasette(“https://datasette.io/content")’ ‘Count available plugins’
Outputs:
There are 154 available plugins.
The output of llm logs -cu shows the tool calls and SQL queries it executed to get that result.
Tags: projects, ai, datasette, generative-ai, llms, llm, llm-tool-use, llm-reasoning, openrouter
AI Summary and Description: Yes
Summary: The text discusses a new release of an LLM plugin for OpenRouter that enhances capabilities related to tool calling and reasoning options. This advancement is significant for professionals in generative AI and LLM security as it provides functionalities to enhance model interactions.
Detailed Description: The provided content outlines the features of the 0.5 release of the LLM plugin for OpenRouter, emphasizing the following key points:
– **Tool Calling Support**: The new release introduces tool calling capabilities, allowing users to interact with any of the 179 tool-enabled models on the OpenRouter platform. This is a pivotal feature for developers looking to integrate AI models with various tools or services.
– **Reasoning Options**: The plugin now supports reasoning efforts, enabling users to specify the complexity of reasoning tasks. For example, users can command the system to ‘prove dogs exist’ with adjustable reasoning effort settings, enhancing the strategic deployment of LLMs.
– **Accessibility**: A segment of the models on OpenRouter can be accessed for free, promoting wider accessibility and experimentation within the generative AI landscape.
– **Practical Examples**:
– The text features commands demonstrating how to install the necessary plugins and execute tool calls, especially highlighting the usage of the llm-tools-datasette plugin with the new Grok 4 Fast model.
– **Output Capabilities**: The command examples illustrate what users can expect as outputs, including counts of available plugins from connected tools and databases.
The text’s details are highly relevant for security and compliance professionals focusing on AI tools and integrations as they touch on enhanced functionality that could lead to new applications or security implications when using AI models.
– **Key Insights**:
– The expansion of tool calling could impact how AI applications interact with various services, thus increasing the complexity of security considerations that need to be addressed.
– Understanding reasoning capabilities is crucial for developing compliant and secure AI applications that meet regulatory requirements.
Overall, the advancements highlighted in this plugin release signify notable progress in LLM capabilities, with implications for both operational usage and associated security practices in AI application development.