OpenLLM-Ro/RoMistral-7b-Instruct-2025-04-23

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Mar 16, 2025License:cc-by-nc-4.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The RoMistral-7b-Instruct-2025-04-23 model, developed by OpenLLM-Ro, is a 7 billion parameter instruction-tuned generative text model specifically designed for the Romanian language. Fine-tuned from Mistral-7B-v0.3, it excels in Romanian natural language tasks, including chat-based interactions and academic benchmarks like XQuAD and STS. This model represents a dedicated open-source effort to provide high-performing LLMs tailored for Romanian linguistic nuances.

Loading preview...

RoMistral-7b-Instruct-2025-04-23 Overview

RoMistral-7b-Instruct-2025-04-23 is a 7 billion parameter instruction-tuned generative text model developed by OpenLLM-Ro, specifically optimized for the Romanian language. It is part of the first open-source initiative to build large language models specialized for Romanian, fine-tuned from the Mistral-7B-v0.3 base model. This model leverages a diverse set of Romanian instruction-tuning datasets, including RoAlpaca, RoDolly, and RoUltraChat, to enhance its performance in conversational and task-oriented scenarios.

Key Capabilities and Performance

  • Romanian Language Specialization: Designed to understand and generate high-quality text in Romanian, addressing the specific linguistic needs of the language.
  • Instruction Following: As an instruct variant, it is well-suited for assistant-like chat and responding to specific instructions.
  • Improved Benchmarks: Demonstrates strong performance across various Romanian benchmarks:
    • Achieves an average score of 54.40 on academic benchmarks, with notable improvements in GSM8k (38.15) and XQuAD F1 (69.11) compared to its predecessors.
    • Scores 6.24 on MT-Bench and 4.36 on RoCulturaBench, indicating robust conversational abilities and cultural understanding within Romanian contexts.
  • Research Focus: Intended primarily for research use in Romanian natural language processing.

Intended Use Cases

  • Research and Development: Ideal for researchers exploring LLM capabilities in Romanian.
  • Assistant-like Chatbots: Suitable for developing conversational AI applications that interact in Romanian.
  • Natural Language Tasks: Can be adapted for various Romanian NLP tasks, leveraging its instruction-tuned nature.

Finetunes
1 models