Guide
What is Function Calling?
A guide to function calling in LLMs: structured tool calls, schemas, APIs, agent workflows, reliability and safety checks.
Function Calling in large language models (LLMs) refers to the structured invocation of external tools, APIs, or workflows through defined schemas to automate complex tasks. It matters now as it enhances AI reliability, safety, and efficiency in real-world applications like tax filings and enterprise engineering. For example, OpenAI's Codex powers self-improving tax agents that reduce errors and streamline workflows, as shown in 12 articles with 12 citations up to May 2026.
Quick Answer
Function calling in LLMs refers to structured tool calls that enable models to interact with APIs and execute workflows. This is increasingly important as AI applications demand reliable and safe interactions with external tools. Recent advancements, such as OpenAI's Codex integration in enterprise environments, highlight the growing significance of function calling in enhancing operational efficiency.
- Evidence base
- 12 filtered articles
- Cited sources
- 12 citations across 5 sources
- Refresh cadence
- Weekly
- Last updated
- Jun 1, 2026
FAQ
What is function calling in LLMs?
Function calling in LLMs refers to structured tool calls that enable models to interact with APIs and execute workflows.
Why is function calling important?
Function calling is crucial for enhancing the reliability and efficiency of AI applications across various sectors.
What recent advancements have been made in function calling?
Recent advancements include the MAVEN framework improving GPT-OSS-120b accuracy from 48% to 71%.
How does OpenAI utilize function calling?
OpenAI utilizes function calling through Codex to automate processes in enterprise environments, enhancing operational efficiency.
Current Read
Function calling in large language models (LLMs) is a critical aspect that allows for structured interactions with external tools and APIs. This capability is essential for enhancing the efficiency and reliability of AI applications across various sectors. For instance, OpenAI's Codex has been instrumental in automating processes in enterprise environments, as seen in partnerships with companies like Cisco and Dell, which aim to integrate AI capabilities securely into their workflows.
Recent developments underscore the importance of function calling in AI. The MAVEN framework has improved the accuracy of the GPT-OSS-120b model from 48% to 71% on the MAVEN-Bench without additional training, showcasing how structured tool calls can enhance model performance. Additionally, OpenAI's recognition as a leader in the 2026 Gartner Magic Quadrant for Enterprise AI Coding Agents emphasizes the growing reliance on function calling for efficient coding solutions in the enterprise landscape.
Key Takeaways
- Function calling enables structured interactions with APIs and tools.
- OpenAI's Codex automates enterprise processes, enhancing efficiency.
- MAVEN improved GPT-OSS-120b accuracy from 48% to 71%.
- OpenAI recognized as a leader in the 2026 Gartner Magic Quadrant.
- Function calling is crucial for reliable AI applications.
Topic Map
Understanding Function Calling
Function calling in LLMs involves structured tool calls that allow models to interact with APIs and execute workflows effectively. This capability is becoming increasingly essential as AI applications require reliable and safe interactions with external systems. The integration of function calling can significantly enhance the operational efficiency of AI solutions.
Recent Advancements in Function Calling
Recent advancements in function calling have shown significant improvements in model performance and application efficiency. For example, the MAVEN framework has demonstrated a substantial increase in accuracy for the GPT-OSS-120b model, highlighting the effectiveness of structured tool calls in enhancing reasoning capabilities. Furthermore, OpenAI's partnerships with companies like Cisco and Dell illustrate the practical applications of function calling in enterprise environments.
Related Guides
What is Tool Use in LLMs?
A guide to LLM tool use: browsing, code execution, APIs, MCP, agents, function calls, guardrails and evaluation.
AI Research Papers This Week
A weekly guide to notable AI research papers across LLMs, agents, inference, robotics, safety and open-source models.
LLM Evaluation and Benchmarks Guide
A guide to LLM evaluation signals: benchmarks, eval methods, reliability, reasoning tests, agents and model comparison.
Source-Linked Articles
MAVEN: Improving Generalization in Agentic Tool Calling
MAVEN (Modular Agentic Verification and Execution Network) enhances reasoning in agentic tool-calling environments, improving GPT-OSS-120b accuracy from 48% to 71% on MAVEN-Bench without extra training. This lightweight framework also remains competitive against proprietary models at a cost ratio of 1/10, highlighting its potential for better compositional reasoning.
arXiv cs.AI · Jun 1, 2026
Building self-improving tax agents with Codex
OpenAI, Thrive, and Crete developed a self-improving tax agent using Codex, which automates tax filings, enhances accuracy, and streamlines workflows. This collaboration aims to reduce human error and improve efficiency in tax processes, significantly benefiting tax professionals and their clients.
OpenAI Blog · May 27, 2026