Key Specifications

Vendorother
Versionstablelm-2-1-6b
Release Date2024-01-19
Context Window4096 tokens
Input Modalitiestext
Output Modalitiestext
LicenseStability AI Community License
Documentationhttps://huggingface.co/models

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 52.7 % 2024-01-19 5-shot view
HUMANEVAL 37.3 pass@1 2024-01-19 view
GSM8K 43.5 % 2024-01-19 0-shot CoT view
MATH 26.4 % 2024-01-19 0-shot CoT view
BBH 40.7 % 2024-01-19 3-shot CoT view
GPQA 15 % 2024-01-19 0-shot view
IFEVAL 44.1 % 2024-01-19 prompt_strict view
ARC 77.7 % 2024-01-19 challenge view
MUSR 26.4 % 2024-01-19 0-shot view
WINOGRANDE 70.3 % 2024-01-19 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 2024-01-19

Compliance

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

StableLM 2 1.6B

Model Overview

Stability AI StableLM 2 1.6B 轻量模型, 4K 上下文, 多语言训练, 适合移动设备部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Other stablelm-2-1-6b 2024-01-19 4K text text Stability AI Community License

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 52.7 % 5-shot
HumanEval 37.3 pass@1
GSM8K (Grade School Math 8K) 43.5 % 0-shot CoT
MATH 26.4 % 0-shot CoT
BBH (BIG-Bench Hard) 40.7 % 3-shot CoT
GPQA 15.0 % 0-shot
IFEval 44.1 % prompt_strict
ARC 77.7 % challenge
MUSR 26.4 % 0-shot
WinoGrande 70.3 % 0-shot

Pricing

Input Output Cache Read Cache Write

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • MMLU 52.7, weak knowledge reasoning.
  • HumanEval 37.3, coding weak.
  • Proprietary, not self-hostable.
  • Context window 4K is limited.

Use Cases

  • General chat and Q&A

References