#2323rd of 23 in Open models
gorilla
Function-calling LLMs, API dataset, and the Berkeley Function Calling Leaderboard
Live demo ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 13.1k
- License
- Apache-2.0
- Last commit
- Mar 2026
- Last release
- Jul 2025
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
Gorilla is a Berkeley research project on getting LLMs to call APIs correctly. The repo holds fine-tuned Gorilla and OpenFunctions models, the APIBench dataset (1,600+ APIs), and the Berkeley Function Calling Leaderboard (BFCL) evaluation code. It also includes GoEx, a runtime for LLM-generated actions, RAFT fine-tuning recipes, and Agent Arena.
Who it is for: Researchers and engineers working on LLM function calling and tool use
Strengths
- Apache 2.0 models, including OpenFunctions-V2 with an OpenAI-compatible function-calling interface
- BFCL covers single-turn, multi-turn and agentic (V4) function calling, with cost and latency metrics
- Includes APIBench dataset and paper evaluation code for reproducing results
- GoEx adds post-facto validation, undo and damage confinement for LLM actions
Weaknesses
- Several unrelated sub-projects in one repo; each has its own setup and README
- Original Gorilla models date from 2023 and may trail newer function-calling models
- Last tagged release is 2025-07-17; roadmap items such as OpenFunctions-v3 remain unchecked
- README gives no RAM or VRAM requirements for running models locally
What it needs
- Models: Gorilla 7B (Hugging Face), Gorilla OpenFunctions-V2, OpenAI-compatible endpoints
Also in Open models
See all 23| Rank | Project | Score |
|---|---|---|
| 1 | llmfitRust tool that checks which LLMs fit your RAM, CPU and GPU | 86 out of 100 |
| 2 | layaSingle-pass multilingual classifier for typed choice, score and yes/no decisions | 83 out of 100 |
| 3 | timesfmPretrained time-series foundation model for zero-shot forecasting from Google Research | 74 out of 100 |
| #4 | unlimited-ocrOCR model that parses long documents and multi-page PDFs in one pass | 69 out of 100 |
| #5 | speechPyTorch toolkit for training and deploying ASR, TTS and speech LLM models | 67 out of 100 |