Key Specifications

Vendorother
Versionnous-capybara-7b
Release Date2023-09-26
Context Window4096 tokens
Input Modalitiestext
Output Modalitiestext
LicenseApache 2.0
Documentationhttps://huggingface.co/models

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 64.4 % 2023-09-26 5-shot view
HUMANEVAL 38.8 pass@1 2023-09-26 view
GSM8K 31 % 2023-09-26 0-shot CoT view
MATH 19.7 % 2023-09-26 0-shot CoT view
BBH 49.6 % 2023-09-26 3-shot CoT view
GPQA 31 % 2023-09-26 0-shot view
IFEVAL 44.2 % 2023-09-26 prompt_strict view
ARC 87 % 2023-09-26 challenge view
MUSR 42.5 % 2023-09-26 0-shot view
WINOGRANDE 77.1 % 2023-09-26 0-shot view

Pricing

Tier Price Currency
Input$0.18 / MtokUSD
Output$0.18 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://huggingface.co/models · as of 2023-09-26

Compliance

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

Nous Capybara 7B

Model Overview

NousResearch Capybara 7B 对话模型, 4K 上下文, 基于 Llama 2 微调, 改进多轮对话能力。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Other nous-capybara-7b 2023-09-26 4K text text Apache 2.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 64.4 % 5-shot
HumanEval 38.8 pass@1
GSM8K (Grade School Math 8K) 31.0 % 0-shot CoT
MATH 19.7 % 0-shot CoT
BBH (BIG-Bench Hard) 49.6 % 3-shot CoT
GPQA 31.0 % 0-shot
IFEval 44.2 % prompt_strict
ARC 87.0 % challenge
MUSR 42.5 % 0-shot
WinoGrande 77.1 % 0-shot

Pricing

Input Output Cache Read Cache Write

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • HumanEval 38.8, coding weak.
  • Proprietary, not self-hostable.
  • Context window 4K is limited.

Use Cases

  • General chat and Q&A

References