CallComply/openchat-3.5-0106-32k

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kTool Calling:SupportedPublished:Jan 11, 2024License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

CallComply/openchat-3.5-0106-32k is a 7 billion parameter instruction-tuned language model developed by OpenChat, featuring a 4096-token context window. This model is optimized for both general chat and coding tasks, and includes a specialized mathematical reasoning mode. It demonstrates strong performance, notably outperforming ChatGPT (March) and Grok-1 on several benchmarks, particularly in coding and mathematical reasoning.

Loading preview...

OpenChat-3.5-0106-32k: A High-Performance 7B Language Model

OpenChat-3.5-0106-32k is a 7 billion parameter instruction-tuned model developed by OpenChat, designed for advanced conversational AI and specialized tasks. This iteration significantly improves upon previous versions, particularly in coding capabilities, and introduces distinct operational modes for enhanced versatility.

Key Capabilities

  • Dual Modes: Features a "Default Mode (GPT4 Correct)" for general chat, coding, and common tasks, and a dedicated "Mathematical Reasoning Mode" tailored for solving complex math problems.
  • Enhanced Coding Performance: Achieves a 15-point improvement in coding benchmarks over its predecessor, OpenChat-3.5, with a HumanEval score of 71.3 and HumanEval+ pass@1 of 65.9.
  • Competitive Benchmarking: Outperforms ChatGPT (March) and Grok-1 on several key benchmarks, including HumanEval, BBH MC, AGIEval, and GSM8K, demonstrating strong overall performance for its size.
  • Experimental Evaluator Support: Includes experimental features for evaluator and feedback capabilities, allowing the model to assess responses based on provided rubrics.

Good For

  • Coding Assistance: Ideal for developers requiring a powerful assistant for code generation, debugging, and understanding.
  • Mathematical Problem Solving: Suitable for applications demanding accurate mathematical reasoning and calculations.
  • General Conversational AI: Effective for chatbots and virtual assistants needing robust general-purpose dialogue capabilities.
  • Research and Evaluation: Useful for researchers exploring model evaluation and feedback mechanisms due to its experimental evaluator features.

Finetunes
2 models