Command R: Complete Benchmark Performance Guide

cohere r 2024-03-11 CC-BY-NC-4.0

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

Key Specifications

Vendorcohere
Versionr
Release Date2024-03-11
Context Window128000 tokens
Input Modalitiestext
Output Modalitiestext
LicenseCC-BY-NC-4.0
Documentationhttps://docs.cohere.com/docs

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 79.3 % 2024-03-11 5-shot view
HUMANEVAL 75.8 pass@1 2024-03-11 — view
GSM8K 86.9 % 2024-03-11 0-shot CoT view
MATH 43.9 % 2024-03-11 0-shot CoT view
BBH 78.4 % 2024-03-11 3-shot CoT view
GPQA 44 % 2024-03-11 0-shot view
IFEVAL 74.1 % 2024-03-11 prompt_strict view
ARC 93.9 % 2024-03-11 challenge view
MUSR 57 % 2024-03-11 0-shot view
WINOGRANDE 84.8 % 2024-03-11 0-shot view

Pricing

Tier Price Currency
Input$0.15 / MtokUSD
Output$0.6 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://cohere.com/pricing · as of 2024-03-11

Compliance

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

Command R

Model Overview

Cohere Command R 经济型 35B 模型, 128K 上下文, 针对 RAG 与工具调用优化, 适合大规模部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Cohere r 2024-03-11 128K text text CC-BY-NC-4.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 79.3 % 5-shot
HumanEval 75.8 pass@1 —
GSM8K (Grade School Math 8K) 86.9 % 0-shot CoT
MATH 43.9 % 0-shot CoT
BBH (BIG-Bench Hard) 78.4 % 3-shot CoT
GPQA 44.0 % 0-shot
IFEval 74.1 % prompt_strict
ARC 93.9 % challenge
MUSR 57.0 % 0-shot
WinoGrande 84.8 % 0-shot

Pricing

Input Output Cache Read Cache Write
— — — —

per million tokens

Strengths

  • GSM8K 86.9, robust math reasoning.
  • Enterprise-grade compliance.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging
  • Enterprise customer support

References