Key Specifications

Vendormistral
Versionmathstral-7b
Release Date2024-07-16
Context Window32768 tokens
Input Modalitiestext
Output Modalitiestext
LicenseApache 2.0
Documentationhttps://docs.mistral.ai/

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU69.1%2024-07-165-shotview
HUMANEVAL49.9pass@12024-07-16view
GSM8K72.4%2024-07-160-shot CoTview
MATH61%2024-07-160-shot CoTview
BBH69.8%2024-07-163-shot CoTview
GPQA37.5%2024-07-160-shotview
IFEVAL63.6%2024-07-16prompt_strictview
ARC91.5%2024-07-16challengeview
MUSR42.3%2024-07-160-shotview
WINOGRANDE71%2024-07-160-shotview

Pricing

TierPriceCurrency
Input$0.18 / MtokUSD
Output$0.18 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://mistral.ai/technology/ · as of 2024-07-16

Compliance

  • Data Residency: EU
  • SOC2: ✓
  • HIPAA: ✗
  • GDPR: ✓
  • ISO 27001: ✓

Mathstral 7B

Modeloverzicht

Mistral Mathstral 7B 数学专用模型, 32K 上下文, 基于 Mistral 7B 微调, 在 MATH 基准上达到 56.6%。

Kernspecificaties

LeverancierVersieReleasedatumContextvensterInvoermodaliteitenUitvoermodaliteitenLicentie
Mistralmathstral-7b2024-07-1632KtexttextApache 2.0

Benchmarkprestaties

BenchmarkScoreEenheidNotities
MMLU (Massive Multitask Language Understanding)69.1%5-shot
HumanEval49.9pass@1
GSM8K (Grade School Math 8K)72.4%0-shot CoT
MATH61.0%0-shot CoT
BBH (BIG-Bench Hard)69.8%3-shot CoT
GPQA37.5%0-shot
IFEval63.6%prompt_strict
ARC91.5%challenge
MUSR42.3%0-shot
WinoGrande71.0%0-shot

Prijzen

InvoerUitvoerCache-lezenCache-schrijven

per miljoen tokens

Sterktes

  • 可靠的通用模型。

Zwaktes

  • HumanEval 49.9,代码能力较弱。
  • 闭源专有模型,不支持自托管。

Gebruiksscenario’s

  • 通用对话与问答

Referenties