Qwen2.5 14B: Complete Benchmark Performance Guide

alibaba 2.5-14b 2024-09-19 Qwen License

In-depth analysis of Qwen2.5 14B performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoralibaba
Version2.5-14b
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 75.4 % 2024-09-19 5-shot view
HUMANEVAL 72 pass@1 2024-09-19 — view
GSM8K 84 % 2024-09-19 0-shot CoT view
MATH 38.8 % 2024-09-19 0-shot CoT view
BBH 72.2 % 2024-09-19 3-shot CoT view
GPQA 44.5 % 2024-09-19 0-shot view
IFEVAL 74.1 % 2024-09-19 prompt_strict view
ARC 94.2 % 2024-09-19 challenge view
MUSR 50.5 % 2024-09-19 0-shot view
WINOGRANDE 84.7 % 2024-09-19 0-shot view

Pricing

Tier Price Currency
Input$0.25 / MtokUSD
Output$0.35 / 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 14B

Model Overview

阿里巴巴 Qwen2.5 14B 开源中型模型, 131K 上下文, 平衡性能与资源, 适合企业级部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Alibaba 2.5-14b 2024-09-19 131K text text Qwen License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 75.4 % 5-shot
HumanEval 72.0 pass@1 —
GSM8K (Grade School Math 8K) 84.0 % 0-shot CoT
MATH 38.8 % 0-shot CoT
BBH (BIG-Bench Hard) 72.2 % 3-shot CoT
GPQA 44.5 % 0-shot
IFEval 74.1 % prompt_strict
ARC 94.2 % challenge
MUSR 50.5 % 0-shot
WinoGrande 84.7 % 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