Qwen2.5 32B: Complete Benchmark Performance Guide

alibaba 2.5-32b 2024-09-19 Qwen License

In-depth analysis of Qwen2.5 32B performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoralibaba
Version2.5-32b
Release Date2024-09-19
Context Window131072 tokens
Input Modalitiestext
Output Modalitiestext
LicenseQwen License
Documentationhttps://qwenlm.github.io/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 78.5 % 2024-09-19 5-shot view
HUMANEVAL 68 pass@1 2024-09-19 — view
GSM8K 76.3 % 2024-09-19 0-shot CoT view
MATH 39.5 % 2024-09-19 0-shot CoT view
BBH 76.8 % 2024-09-19 3-shot CoT view
GPQA 39.4 % 2024-09-19 0-shot view
IFEVAL 73.5 % 2024-09-19 prompt_strict view
ARC 93.9 % 2024-09-19 challenge view
MUSR 54.9 % 2024-09-19 0-shot view
WINOGRANDE 81.9 % 2024-09-19 0-shot view

Pricing

Tier Price Currency
Input$0.35 / MtokUSD
Output$0.45 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://help.aliyun.com/zh/model-studio/getting-started/models · as of 2024-09-19

Compliance

  • Data Residency: CN
  • SOC2: ✗
  • HIPAA: ✗
  • GDPR: ✗
  • ISO 27001: ✗

Qwen2.5 32B

Model Overview

阿里巴巴 Qwen2.5 32B 开源中端模型, 131K 上下文, 性能接近 72B, 适合中等规模部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Alibaba 2.5-32b 2024-09-19 131K text text Qwen License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 78.5 % 5-shot
HumanEval 68.0 pass@1 —
GSM8K (Grade School Math 8K) 76.3 % 0-shot CoT
MATH 39.5 % 0-shot CoT
BBH (BIG-Bench Hard) 76.8 % 3-shot CoT
GPQA 39.4 % 0-shot
IFEval 73.5 % prompt_strict
ARC 93.9 % challenge
MUSR 54.9 % 0-shot
WinoGrande 81.9 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging

References