Qwen2.5-Turbo: Advanced Language Model
date
Nov 24, 2024
damn
language
en
status
Published
type
Products
image
https://www.ai-damn.com/1732419397128-202411211401179308.jpg
slug
qwen2-5-turbo-advanced-language-model-1732419403257
tags
AI Language Model
Text Processing
Alibaba
Natural Language Processing
High Performance
summary
Qwen2.5-Turbo is an innovative language model designed by Alibaba's team, optimized for efficient long text processing. It supports a context of up to 1 million tokens, making it ideal for handling extensive text data while maintaining high performance and cost-effectiveness.
Product Introduction
Qwen2.5-Turbo is an advanced language model developed by Alibaba, tailored for efficient processing of long texts. With the capability to manage a context of up to 1 million tokens, it excels in both long and short text processing, offering remarkable accuracy and cost-effectiveness for various applications.
Key Features
- Supports a context length of up to 1 million tokens.
- Achieved 100% accuracy in the 1M-token Passkey Retrieval task.
- Scored 93.1 in the RULER long text evaluation benchmarks, outperforming GPT-4.
- Integrates sparse attention mechanisms, reducing the time to generate the first token from 4.9 minutes to 68 seconds.
- Processing cost is competitive at only 0.3 yuan per million tokens.
- High performance in short text processing comparable to GPT-4o-mini.
Product Data
- Model Name: Qwen2.5-Turbo
- Context Length: Up to 1 million tokens (approx. 1 million English words or 1.5 million Chinese characters)
- Accuracy Rate: 100% in Passkey Retrieval task
- RULER Benchmark Score: 93.1
- Cost: 0.3 yuan per million tokens processed