Qwen1.5 110B: Complete Benchmark Performance Guide

alibaba 1.5-110b 2024-02-25 Qwen License

In-depth analysis of Qwen1.5 110B performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendoralibaba
Version1.5-110b
Release Date2024-02-25
Context Window32768 tokens
Input Modalitiestext
Output Modalitiestext
LicenseQwen License
Documentationhttps://qwenlm.github.io/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 86.1 % 2024-02-25 5-shot view
HUMANEVAL 76.5 pass@1 2024-02-25 — view
GSM8K 80.6 % 2024-02-25 0-shot CoT view
MATH 54.1 % 2024-02-25 0-shot CoT view
BBH 82.3 % 2024-02-25 3-shot CoT view
GPQA 45.9 % 2024-02-25 0-shot view
IFEVAL 83.4 % 2024-02-25 prompt_strict view
ARC 95.7 % 2024-02-25 challenge view
MUSR 63.9 % 2024-02-25 0-shot view
WINOGRANDE 84 % 2024-02-25 0-shot view

Pricing

Tier Price Currency
Input$0.7 / MtokUSD
Output$1.2 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://help.aliyun.com/zh/model-studio/getting-started/models · as of 2024-02-25

Compliance

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

Qwen1.5 110B

Model Overview

阿里巴巴 Qwen1.5 110B 开源超大模型, 32K 上下文, 在中文与多语言基准上表现突出。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Alibaba 1.5-110b 2024-02-25 32K text text Qwen License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 86.1 % 5-shot
HumanEval 76.5 pass@1 —
GSM8K (Grade School Math 8K) 80.6 % 0-shot CoT
MATH 54.1 % 0-shot CoT
BBH (BIG-Bench Hard) 82.3 % 3-shot CoT
GPQA 45.9 % 0-shot
IFEval 83.4 % prompt_strict
ARC 95.7 % challenge
MUSR 63.9 % 0-shot
WinoGrande 84.0 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • MMLU score 86.1, strong knowledge reasoning.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging

References