GSoC Participation History
Technologies
Topics
Past Projects
Evaluate Gemini on an Open-Source Benchmark
This project will evaluate Google's Gemini (2.5 Pro, 2.0 Flash) and Gemma (3, CodeGemma) models by extending multiple open-source coding benchmarks...
Implementing Web Interface For Fine-Tuning Gemma Models Using Gradio
This project is about building a user-friendly web interface using Gradio that makes fine-tuning Gemma models simple and accessible for everyone—even...
Improving Gemini Documentation for Open Source Model Providers Promptfoo and Weights & Biases
The project will close the documentation and evaluation gap for Google’s Gemini models by contributing structured onboarding guides and benchmarking...
Evaluate Gemini on an Open-Source Benchmark
This project aims to rigorously evaluate Google’s Gemini multimodal model on real-world tasks. We plan to address the limitations of current...
Enhance Gemini API Integrations in OSS Agents Tools
This project will elevate Gemini API support across widely used open-source agent frameworks like LangChain, LlamaIndex, CrewAI, and...
Gemma Chat Gradio Demo
A majority of the Gemma chat applications on Hugging Face Spaces do not allow the user to adjust generation settings or system prompts, giving the...
Unified Gemini Example Cookbook: Migrating and Modernizing Open-Source Learning Resources
This proposal aims to upgrade and expand existing open-source tutorials and examples to support the new unified Gemini SDKs for JavaScript/TypeScript...
Develop a Gemini Workspace in Postman
This project aims to create a Gemini Workspace in Postman for interacting with the Gemini API’s and providing a central hub for exploration ,...
EchoGem – Teaching Gemini to Think in Batches by Prioritizing What Matters
EchoGem introduces a novel batching engine designed to answer multiple questions about the same source parallelly to reduce response times heavily....
Gemma Model Fine-tuning UI
Gemma is a lightweight, open-source large language model by Google DeepMind. This project aims to build an intuitive web interface for fine-tuning...
Creation of a Creative Thinking Benchmark
The goal of this project is to develop a multi-modal and open-source benchmark with which to evaluate Gemini 2.0. Open-source benchmarks are an...
Modernizing Gemini SDK Learning Resources: Migration, Tutorials, and Library Updates
This proposal modernizes Gemini SDK learning resources by migrating outdated examples to JavaScript / TypeScript, developing new tutorials, and...
Develop a Gemini Workspace in Postman
This project aims to develop a comprehensive Postman workspace for Google's Gemini API, providing a central hub for exploration, integration, and...
Comprehensive Benchmark Suite for Evaluating Gemma Models
This project will develop a robust, extensible benchmarking suite designed for Google's Gemma models to address the lack of standardized evaluation...
Enhancing Gemini Integrations in Open-Source Agent Frameworks
Google DeepMind’s Gemini is a powerful multimodal model that excels at language comprehension, visual processing, and tool interaction. However,...
HALO: Hierarchical Abstraction for Longform Optimization
HALO is an innovative open-source framework designed to optimize Gemini API usage for long-context video analysis. By leveraging hierarchical...
Web Interface for Accessible, Configurable Gemma Open Models Fine-Tuning Workflows
Fine-tuning LLMs like Gemma requires deep technical understanding and complex configuration, hindering rapid prototyping and broader application....
Gemini API Developer Workspace in Postman
This project proposes the creation of a comprehensive Postman Workspace tailored for Google’s Gemini API suite. It will offer developers a robust,...
VS Code Extension to assist with coding powered by Gemini APIs
A VS Code (or JetBrains) extension to provide AI-powered coding assistance using Google’s Gemini API. This tool enhances the developer experience by...
Develop a Gemini Workspace in Postman
The goal of this project is to create a developer-friendly Postman Workspace for interacting with the Gemini API. This workspace will serve as a...
Gemma Garage: Leveraging Gemma 3 to democratize LLM Fine-tuning
This proposal aims to develop the Gemma LLM Garage, a full-stack interface to manage datasets and fine-tune Gemma models. Its main goal is to...
Enhanced Benchmark for Evaluating Intuitive Physics Understanding in Gemma Multimodal Models
This project aims to develop a more rigorous and focused evaluation testbed than that used by Garrido et al. (2025), with the specific goal of...
ATIA: A BENCHMARK FOR ADVERSARIAL TOOL INFILTRATION IN AGENTS
As multimodal agents become increasingly integrated into real-world applications, ensuring their safe and reliable tool-use behavior is paramount. We...
Xarray-JAX Integration Library
This project aims to develop a Python library to support Xarray data (labeled multi-dimensional array library supported by Deepmind) with JAX...
Enhance Gemini API Integrations in OSS Agents Tools (DeepMind)
This project aims to enhance Gemini API integrations in open-source software (OSS) agent tools, specifically LangChain and LlamaIndex, by addressing...
Streamline experiment execution and improve report UI for OSS-Fuzz-Gen
OSS-Fuzz-Gen, a framework using LLMs for fuzz target generation and evaluation by Google, currently has a basic experiment report UI alongside a...
Highly Cost-Efficient Fine-Tuning of Gemma 3 to Develop Test-Time Scaling for Visual/Spatial Tasks.
Fine-tuning MLLMs to achieve test-time scaling and strong reasoning performance for visual/spatial tasks can be very expensive because of the...
SciResearchBench: A Multimodal Benchmark for Scientific Reasoning and Discovery
Scientific discovery fundamentally relies on integrating and reasoning over multimodal information—text, diagrams, plots, spectra, microscopy images,...
Batch Prediction with Long Context and Context Caching Code Sample
The aim of this project is to develop a code sample in Python that demonstrates some of the key capabilities of Google's Gemini APIs in regards to...
Reproducibility as Accuracy (RaA) Benchmark
Reproducibility as Accuracy (RaA) is a benchmark which aims to evaluate how effective multimodal AI systems are in preserving information fidelity...
Streamlining Gemini API Development: Building a Postman Workspace for Gemini API’s
When attempting to utilize Google's Gemini APIs, developers face a steep learning curve. The scattered documentation, complex parameter...
Gemma Function-Calling Sandbox for Real-World Applications
AI models with function-calling capabilities can be powerful tools for solving real-world problems, but experimenting with them isn’t always...
GemmaEval: A Comprehensive Automated Benchmark Suite for Gemma
GemmaEval is a comprehensive, automated benchmarking framework designed to evaluate, visualize, and compare Gemma language models against other open...
Open-Source Multimodal Benchmarks and Adversarial Robustness Testing for Gemini 2.X models
This project aims to advance the evaluation framework for Google’s Gemini 2.0 and Gemma 3-27B multimodal models by integrating a diverse set of...
Batch Prediction Framework: Long Context and Context Caching for Video Analysis
This project develops an efficient framework for analyzing educational video content using the Gemini API. The approach combines optimized batch...
Improve Evals Documentation for the Gemini APIs
Current manual evaluation methods for LLM-based AI applications are unsustainable and resource-intensive. While using LLMs as judges offers a...
Creating New Agent Architectures for Concordia
The goal of this project is to help strengthen the Concordia framework by developing and open-sourcing a collection of new language model agent...
Enhance Gemini Support in Open-Source Extensions (Continue.dev/Aider-like)
This project aims to enhance the integration of Google's Gemini AI into open-source IDE extensions, focusing on developing and improving user...
Open-source Gemini Example Apps
The Gemini Cookbook is a set of sample applications and tutorials illustrating different functionalities of the Gemini APIs. The intent of this...
Batch Prediction with Long Context and Context Caching
This project aims to build a production-grade, open-source code sample that showcases batch question answering over long-context inputs using Google...
Exploring & Extending Function Calling in Gemma
This project aims to: (a) investigate to explore technical possibilities, enhance specifications, and find applications for specific use cases and...
Multimodal Intelligence: Supercharging Agents with Gemini
This project addresses critical gaps in Gemini API integration across leading agent frameworks (LangChain, LlamaIndex, CrewAI, Composio), where...
Evaluate Gemini (Gemma) on an Open-Source Benchmark
This project aims to design a robust, reproducible benchmark for evaluating the Gemini family of multimodal large language models (MLLMs), such as...
Gemma Model Function Calling Exploration
Enable and document best practices for function calling in Gemma models, similar to the structured tool-use capabilities . In essence, function...
Self-Contained OSS-Fuzz Module for Researchers
This project aims to develop a standalone Python SDK that provides researchers with a streamlined and well-documented API for interacting with...
Frequently Asked Questions
Google DeepMind | GSoC Org Profile & Stats - Learn about Google DeepMind's involvement in Google Summer of Code (GSoC), their technologies, detailed reports.
Participation
Projects
Top Programming Languages
Project Difficulty Distribution
No difficulty data available
Built with ❤️ for the Google Summer of Code (GSoC) open source community | © 2026 GSoCOrganizationsGuide