Qwen2 7B: Complete Benchmark Performance Guide

alibaba 2-7b 2024-06-07 Qwen License

In-depth analysis of Qwen2 7B performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoralibaba
Version2-7b
Release Date2024-06-07
Context Window131072 tokens
Input Modalitiestext
Output Modalitiestext
LicenseQwen License
Documentationhttps://qwenlm.github.io/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 63.4 % 2024-06-07 5-shot view
HUMANEVAL 67 pass@1 2024-06-07 — view
GSM8K 73.2 % 2024-06-07 0-shot CoT view
MATH 40.7 % 2024-06-07 0-shot CoT view
BBH 63.3 % 2024-06-07 3-shot CoT view
GPQA 31.6 % 2024-06-07 0-shot view
IFEVAL 61.8 % 2024-06-07 prompt_strict view
ARC 86.8 % 2024-06-07 challenge view
MUSR 43.9 % 2024-06-07 0-shot view
WINOGRANDE 77.3 % 2024-06-07 0-shot view

Pricing

Tier Price Currency
Input$0.15 / MtokUSD
Output$0.15 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://help.aliyun.com/zh/model-studio/getting-started/models · as of 2024-06-07

Compliance

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

Qwen2 7B

Model Overview

阿里巴巴 Qwen2 7B 经济型开源模型, 131K 上下文, 7B 参数, 性能超越 Llama 2 13B, 适合本地部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Alibaba 2-7b 2024-06-07 131K text text Qwen License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 63.4 % 5-shot
HumanEval 67.0 pass@1 —
GSM8K (Grade School Math 8K) 73.2 % 0-shot CoT
MATH 40.7 % 0-shot CoT
BBH (BIG-Bench Hard) 63.3 % 3-shot CoT
GPQA 31.6 % 0-shot
IFEval 61.8 % prompt_strict
ARC 86.8 % challenge
MUSR 43.9 % 0-shot
WinoGrande 77.3 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • General chat and Q&A

References