Key Specifications

Vendormeta
Versioncode-llama-13b
Release Date2023-08-24
Context Window16000 tokens
Input Modalitiestext
Output Modalitiestext
LicenseLlama 2 Community License
Documentationhttps://llama.meta.com/docs/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 65.9 % 2023-08-24 5-shot view
HUMANEVAL 69.5 pass@1 2023-08-24 view
GSM8K 55.9 % 2023-08-24 0-shot CoT view
MATH 37.8 % 2023-08-24 0-shot CoT view
BBH 71 % 2023-08-24 3-shot CoT view
GPQA 36.4 % 2023-08-24 0-shot view
IFEVAL 50.1 % 2023-08-24 prompt_strict view
ARC 87.8 % 2023-08-24 challenge view
MUSR 47.3 % 2023-08-24 0-shot view
WINOGRANDE 70.4 % 2023-08-24 0-shot view

Pricing

Tier Price Currency
Input$0.3 / MtokUSD
Output$0.3 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://ai.meta.com/blog/ · as of 2023-08-24

Compliance

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

Code Llama 13B

Model Overview

Meta Code Llama 13B 代码专用开源模型, 16K 上下文, 13B 参数, 适合中等规模代码任务部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Meta code-llama-13b 2023-08-24 16K text text Llama 2 Community License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 65.9 % 5-shot
HumanEval 69.5 pass@1
GSM8K (Grade School Math 8K) 55.9 % 0-shot CoT
MATH 37.8 % 0-shot CoT
BBH (BIG-Bench Hard) 71.0 % 3-shot CoT
GPQA 36.4 % 0-shot
IFEval 50.1 % prompt_strict
ARC 87.8 % challenge
MUSR 47.3 % 0-shot
WinoGrande 70.4 % 0-shot

Pricing

Input Output Cache Read Cache Write

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • Proprietary, not self-hostable.
  • Context window 16K is limited.

Use Cases

  • Code generation and debugging

References