Rubra
Click to visit website
About
Rubra is a collection of open-weight, tool-calling LLMs designed to enhance top open-source large language models. It adds deterministic tool-calling capability, making these models ideal for agentic use cases. Rubra models are post-trained and use methods to teach new skills while mitigating catastrophic forgetting. They extend popular inferencing projects like llama.cpp and vLLM for easy local use, providing OpenAI-compatible tool-calling. Rubra excels in complex, multi-step function calls, outperforming original models and other fine-tunes in this area. The models are trained on a high-quality dataset of over 1 million conversations and tool calls, with thousands of A100/H100 GPU hours. Rubra models are published under the same license as their parent models, while Rubra code is Apache 2.0 licensed.
Platform
Task
Features
• free demo available on huggingface spaces
• ideal for agentic use cases
• supports complex, multi-step function calls ('chain of function')
• extends popular inferencing projects (llama.cpp, vllm) for local use
• mitigates catastrophic forgetting during enhancement
• deterministic tool-calling capability
• enhances top open-source llms (meta-llama-3, gemma, mistral, phi-3, qwen2)
• open-weight, tool-calling llms
FAQs
Who are the intended users of Rubra?
Rubra models are for anyone looking to use open source LLMs with native function calling support, which yields superior results in local LLMs when compared to prompting models to return tool calls.
Why use Rubra models over Llama, Mistral, or other popular fine tunes like Hermes Pro or Gorilla OpenFunctions?
Rubra models are capable of complex, multi-step, function calls (chain of function) and enhance popular open source instruct LLMs while retaining their original capabilities.
Mistral-7B-Instruct-v0.3 has tool calling capability, so why use Rubra enhanced Mistral-7B-Instruct-v0.3
Rubra enhanced Mistral-7B-Instruct-v0.3 is capable of complex tool calling that the original model falls short of, demonstrating superior multi-step function chaining.
vLLM has tool calling capability, so why use Rubra?
Rubra enhanced models and custom vLLM give the LLM full discretion on when to make a tool call or reply with an assistant message, unlike vLLM which expects tool calls.
How were Rubra models trained?
Models were trained on a high-quality tool calling dataset (1M+ conversations) using thousands of A100/H100 GPU hours, with iterative training techniques for fast convergence.
Why do the benchmark results differ from the ones found in parent model cards?
Benchmark results differ due to constant improvements in evaluation tools (LM Evaluation Harness, FastChat LLM Judge) and updates to judging models, but Rubra does not game benchmarks.
How did you construct the training set for Rubra models?
The training set was constructed using the 'chain of function' concept, focusing on multi-step, consecutive function calls to achieve complex goals, exemplified by customer feedback processing.
Pricing Plans
Free
Free Plan• Open-weight, tool-calling LLMs
• Access to enhanced models
• Demo via Huggingface Spaces
• Local deployment with extended inferencing tools
Job Opportunities
There are currently no job postings for this AI tool.
Ratings & Reviews
No ratings available yet. Be the first to rate this tool!
Featured Tools
adly.news
adly.news is a free platform that simplifies newsletter advertising, connecting businesses with engaged audiences through ad slots, offering bidding, negotiation, and messaging.
View Details
Gemini Watermark Remover
Gemini Watermark Remover is a client-side tool designed to remove hidden SynthID and other embedded watermarks from your AI-generated images, preserving quality.
View Details
Infatuated.AI
Infatuated.AI is an AI companion platform allowing users to chat, roleplay, and build personalized relationships with AI girlfriends and boyfriends, offering emotional support and secure fantasy sharing.
View Details
ImgGen
ImgGen is the free AI editor that edits photos and turns images into videos in seconds, offering instant creativity all in one place.
View Details
Nano Banana
Nano Banana is a state-of-the-art AI model that revolutionizes text-based image editing and generation with unmatched multi-image fusion and natural language understanding.
View Details
Macaron
Macaron is the world’s first personal AI agent designed to help you live better by focusing on happiness, health, and freedom, unlike typical productivity tools.
View Details
VISBOOM
Visboom is the all-in-one AI fashion content creation platform, enabling brands and e-commerce sellers to generate on-model photoshoots and visual assets quickly.
View Details
Banana AI
Banana AI is an advanced AI photo editor powered by Google’s Nano Banana technology (Gemini 2.5 Flash Image), enabling effortless image editing, restyling, and transformation with simple text prompts.
View Details
twainGPT
twainGPT is a humanizer that transforms any AI-generated text into undetectable, human-like content, trusted by over 2.3 million users.
View Details