Key Specifications

Vendorother
Versionzephyr-orpo-141b-alpha
Release Date2024-03-22
Context Window4096 tokens
Input Modalitiestext
Output Modalitiestext
LicenseApache 2.0
Documentationhttps://huggingface.co/models

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 79.5 % 2024-03-22 5-shot view
HUMANEVAL 74.7 pass@1 2024-03-22 view
GSM8K 80.6 % 2024-03-22 0-shot CoT view
MATH 36 % 2024-03-22 0-shot CoT view
BBH 70.3 % 2024-03-22 3-shot CoT view
GPQA 36.8 % 2024-03-22 0-shot view
IFEVAL 69.7 % 2024-03-22 prompt_strict view
ARC 91.4 % 2024-03-22 challenge view
MUSR 57.7 % 2024-03-22 0-shot view
WINOGRANDE 79.7 % 2024-03-22 0-shot view

Pricing

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

Source: https://huggingface.co/models · as of 2024-03-22

Compliance

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

Zephyr ORPO 141B Alpha

Model Overview

HuggingFaceH4 Zephyr ORPO 141B Alpha, 4K 上下文, 基于 Mixtral 8x22B ORPO 微调, 性能接近 GPT-4。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Other zephyr-orpo-141b-alpha 2024-03-22 4K text text Apache 2.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 79.5 % 5-shot
HumanEval 74.7 pass@1
GSM8K (Grade School Math 8K) 80.6 % 0-shot CoT
MATH 36.0 % 0-shot CoT
BBH (BIG-Bench Hard) 70.3 % 3-shot CoT
GPQA 36.8 % 0-shot
IFEval 69.7 % prompt_strict
ARC 91.4 % challenge
MUSR 57.7 % 0-shot
WinoGrande 79.7 % 0-shot

Pricing

Input Output Cache Read Cache Write

per million tokens

Strengths

  • Mixture-of-Experts architecture.

Weaknesses

  • Proprietary, not self-hostable.
  • Context window 4K is limited.

Use Cases

  • Code generation and debugging

References