GLM-4 Air: Complete Benchmark Performance Guide

other glm-4-air 2024-08-12 Proprietary

In-depth analysis of GLM-4 Air performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendorother
Versionglm-4-air
Release Date2024-08-12
Context Window131072 tokens
Input Modalitiestext
Output Modalitiestext
LicenseProprietary
Documentationhttps://huggingface.co/models

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 78.9 % 2024-08-12 5-shot view
HUMANEVAL 68.3 pass@1 2024-08-12 — view
GSM8K 82.3 % 2024-08-12 0-shot CoT view
MATH 40.6 % 2024-08-12 0-shot CoT view
BBH 79 % 2024-08-12 3-shot CoT view
GPQA 31.3 % 2024-08-12 0-shot view
IFEVAL 75.7 % 2024-08-12 prompt_strict view
ARC 92.3 % 2024-08-12 challenge view
MUSR 46.7 % 2024-08-12 0-shot view
WINOGRANDE 78.1 % 2024-08-12 0-shot view

Pricing

Tier Price Currency
Input$0.1 / MtokUSD
Output$0.1 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://huggingface.co/models · as of 2024-08-12

Compliance

  • Data Residency: self-host
  • SOC2: ✗
  • HIPAA: ✗
  • GDPR: ✗
  • ISO 27001: ✗

GLM-4 Air

Model Overview

清华智谱 GLM-4 Air 经济型模型, 131K 上下文, 速度快成本低, 适合大规模部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Other glm-4-air 2024-08-12 131K text text Proprietary

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 78.9 % 5-shot
HumanEval 68.3 pass@1 —
GSM8K (Grade School Math 8K) 82.3 % 0-shot CoT
MATH 40.6 % 0-shot CoT
BBH (BIG-Bench Hard) 79.0 % 3-shot CoT
GPQA 31.3 % 0-shot
IFEval 75.7 % prompt_strict
ARC 92.3 % challenge
MUSR 46.7 % 0-shot
WinoGrande 78.1 % 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