Key Specifications

Vendoralibaba
Version2-1-5b
Release Date2024-06-07
Context Window32768 tokens
Input Modalitiestext
Output Modalitiestext
LicenseQwen License
Documentationhttps://qwenlm.github.io/

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU40.5%2024-06-075-shotview
HUMANEVAL47.2pass@12024-06-07view
GSM8K36.2%2024-06-070-shot CoTview
MATH25.4%2024-06-070-shot CoTview
BBH51.5%2024-06-073-shot CoTview
GPQA22.1%2024-06-070-shotview
IFEVAL57.4%2024-06-07prompt_strictview
ARC85.4%2024-06-07challengeview
MUSR27.8%2024-06-070-shotview
WINOGRANDE62.9%2024-06-070-shotview

Pricing

TierPriceCurrency
Input$0.05 / MtokUSD
Output$0.05 / 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 1.5B

Aperçu du modèle

阿里巴巴 Qwen2 1.5B 轻量开源模型, 32K 上下文, 1.5B 参数, 适合边缘设备与移动部署。

Spécifications principales

FournisseurVersionDate de sortieFenêtre de contexteModalités d’entréeModalités de sortieLicence
Alibaba2-1-5b2024-06-0732KtexttextQwen License

Performance aux benchmarks

BenchmarkScoreUnitéNotes
MMLU (Massive Multitask Language Understanding)40.5%5-shot
HumanEval47.2pass@1
GSM8K (Grade School Math 8K)36.2%0-shot CoT
MATH25.4%0-shot CoT
BBH (BIG-Bench Hard)51.5%3-shot CoT
GPQA22.1%0-shot
IFEval57.4%prompt_strict
ARC85.4%challenge
MUSR27.8%0-shot
WinoGrande62.9%0-shot

Tarification

EntréeSortieLecture cacheÉcriture cache

par million de jetons

Forces

  • 可靠的通用模型。

Faiblesses

  • MMLU 仅 40.5,知识推理偏弱。
  • HumanEval 47.2,代码能力较弱。
  • 闭源专有模型,不支持自托管。

Cas d’usage

  • 通用对话与问答

Références