Mistral Medium: Complete Benchmark Performance Guide

mistral medium 2023-12-11 Apache 2.0

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

Key Specifications

Vendormistral
Versionmedium
Release Date2023-12-11
Context Window32000 tokens
Input Modalitiestext
Output Modalitiestext
LicenseApache 2.0
Documentationhttps://docs.mistral.ai/

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 82.9 % 2023-12-11 5-shot view
HUMANEVAL 78 pass@1 2023-12-11 — view
GSM8K 82.5 % 2023-12-11 0-shot CoT view
MATH 54.9 % 2023-12-11 0-shot CoT view
BBH 83.4 % 2023-12-11 3-shot CoT view
GPQA 48.7 % 2023-12-11 0-shot view
IFEVAL 77.4 % 2023-12-11 prompt_strict view
ARC 94.6 % 2023-12-11 challenge view
MUSR 64.3 % 2023-12-11 0-shot view
WINOGRANDE 82.4 % 2023-12-11 0-shot view

Pricing

Tier Price Currency
Input$2.7 / MtokUSD
Output$8.1 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://mistral.ai/technology/ · as of 2023-12-11

Compliance

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

Mistral Medium

Model Overview

Mistral AI Medium 中端模型, 32K 上下文, 平衡性能与成本, 适合企业级通用任务。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Mistral medium 2023-12-11 32K text text Apache 2.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 82.9 % 5-shot
HumanEval 78.0 pass@1 —
GSM8K (Grade School Math 8K) 82.5 % 0-shot CoT
MATH 54.9 % 0-shot CoT
BBH (BIG-Bench Hard) 83.4 % 3-shot CoT
GPQA 48.7 % 0-shot
IFEval 77.4 % prompt_strict
ARC 94.6 % challenge
MUSR 64.3 % 0-shot
WinoGrande 82.4 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • MMLU score 82.9, strong knowledge reasoning.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging

References