Key Specifications

Vendorother
Versionsonar-huge
Release Date2024-08-13
Context Window127072 tokens
Input Modalitiestext
Output Modalitiestext
LicenseProprietary
Documentationhttps://huggingface.co/models

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU81%2024-08-135-shotview
HUMANEVAL75.8pass@12024-08-13view
GSM8K90.4%2024-08-130-shot CoTview
MATH46.8%2024-08-130-shot CoTview
BBH78.6%2024-08-133-shot CoTview
GPQA46.7%2024-08-130-shotview
IFEVAL78.7%2024-08-13prompt_strictview
ARC94.8%2024-08-13challengeview
MUSR55.4%2024-08-130-shotview
WINOGRANDE87.7%2024-08-130-shotview

Pricing

TierPriceCurrency
Input$2 / MtokUSD
Output$2 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://huggingface.co/models · as of 2024-08-13

Compliance

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

Sonar Huge

Descripción del modelo

Perplexity Sonar Huge 旗舰在线 RAG 模型, 127K 上下文, 基于 Llama 3.1 405B 微调, 推理能力最强。

Especificaciones principales

ProveedorVersiónFecha de lanzamientoVentana de contextoModalidades de entradaModalidades de salidaLicencia
Othersonar-huge2024-08-13127KtexttextProprietary

Rendimiento en benchmarks

BenchmarkPuntuaciónUnidadNotas
MMLU (Massive Multitask Language Understanding)81.0%5-shot
HumanEval75.8pass@1
GSM8K (Grade School Math 8K)90.4%0-shot CoT
MATH46.8%0-shot CoT
BBH (BIG-Bench Hard)78.6%3-shot CoT
GPQA46.7%0-shot
IFEval78.7%prompt_strict
ARC94.8%challenge
MUSR55.4%0-shot
WinoGrande87.7%0-shot

Precios

EntradaSalidaLectura cachéEscritura caché

por millón de tokens

Fortalezas

  • MMLU score 81.0, strong knowledge reasoning.
  • GSM8K 90.4, robust math reasoning.
  • 企业级合规认证。

Debilidades

  • 闭源专有模型,不支持自托管。

Casos de uso

  • 代码生成与调试
  • 企业客户支持

Referencias