Mistral 7B v0.3: Complete Benchmark Performance Guide

mistral 7b-v0.3 2024-05-22 Apache 2.0

In-depth analysis of Mistral 7B v0.3 performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendormistral
Version7b-v0.3
Release Date2024-05-22
Context Window32768 tokens
Input Modalitiestext
Output Modalitiestext
LicenseApache 2.0
Documentationhttps://docs.mistral.ai/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 72.6 % 2024-05-22 5-shot view
HUMANEVAL 68.8 pass@1 2024-05-22 — view
GSM8K 69.4 % 2024-05-22 0-shot CoT view
MATH 26.7 % 2024-05-22 0-shot CoT view
BBH 60.5 % 2024-05-22 3-shot CoT view
GPQA 31.2 % 2024-05-22 0-shot view
IFEVAL 60.9 % 2024-05-22 prompt_strict view
ARC 88.9 % 2024-05-22 challenge view
MUSR 48.1 % 2024-05-22 0-shot view
WINOGRANDE 72.7 % 2024-05-22 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://mistral.ai/technology/ · as of 2024-05-22

Compliance

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

Mistral 7B v0.3

Model Overview

Mistral 7B v0.3 开源模型, 32K 上下文, 改进 v0.2, 扩展词汇表至 32768, 适合本地部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Mistral 7b-v0.3 2024-05-22 32K text text Apache 2.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 72.6 % 5-shot
HumanEval 68.8 pass@1 —
GSM8K (Grade School Math 8K) 69.4 % 0-shot CoT
MATH 26.7 % 0-shot CoT
BBH (BIG-Bench Hard) 60.5 % 3-shot CoT
GPQA 31.2 % 0-shot
IFEval 60.9 % prompt_strict
ARC 88.9 % challenge
MUSR 48.1 % 0-shot
WinoGrande 72.7 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • Reliable general-purpose model.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • General chat and Q&A

References