Command R7B: Complete Benchmark Performance Guide

cohere r7b 2024-08-13 CC-BY-NC-4.0

In-depth analysis of Command R7B performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendorcohere
Versionr7b
Release Date2024-08-13
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 64.9 % 2024-08-13 5-shot view
HUMANEVAL 50.1 pass@1 2024-08-13 — view
GSM8K 69.7 % 2024-08-13 0-shot CoT view
MATH 40.6 % 2024-08-13 0-shot CoT view
BBH 66.1 % 2024-08-13 3-shot CoT view
GPQA 30.9 % 2024-08-13 0-shot view
IFEVAL 67.2 % 2024-08-13 prompt_strict view
ARC 90.7 % 2024-08-13 challenge view
MUSR 47 % 2024-08-13 0-shot view
WINOGRANDE 74.1 % 2024-08-13 0-shot view

Pricing

Tier Price Currency
Input$0.0375 / MtokUSD
Output$0.0375 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://cohere.com/pricing · as of 2024-08-13

Compliance

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

Command R7B

Model Overview

Cohere Command R7B 轻量开源模型, 128K 上下文, 7B 参数, 为 RAG 与工具调用优化, 适合边缘部署。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Cohere r7b 2024-08-13 128K text text CC-BY-NC-4.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 64.9 % 5-shot
HumanEval 50.1 pass@1 —
GSM8K (Grade School Math 8K) 69.7 % 0-shot CoT
MATH 40.6 % 0-shot CoT
BBH (BIG-Bench Hard) 66.1 % 3-shot CoT
GPQA 30.9 % 0-shot
IFEval 67.2 % prompt_strict
ARC 90.7 % challenge
MUSR 47.0 % 0-shot
WinoGrande 74.1 % 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