Key Specifications

Vendorcohere
Versionnightly
Release Date2024-01-01
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 85.2 % 2024-01-01 5-shot view
HUMANEVAL 74.7 pass@1 2024-01-01 view
GSM8K 87.2 % 2024-01-01 0-shot CoT view
MATH 49.2 % 2024-01-01 0-shot CoT view
BBH 82.7 % 2024-01-01 3-shot CoT view
GPQA 45.5 % 2024-01-01 0-shot view
IFEVAL 80.8 % 2024-01-01 prompt_strict view
ARC 95.9 % 2024-01-01 challenge view
MUSR 57.5 % 2024-01-01 0-shot view
WINOGRANDE 80.8 % 2024-01-01 0-shot view

Pricing

Tier Price Currency
Input$2.5 / MtokUSD
Output$10 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://cohere.com/pricing · as of 2024-01-01

Compliance

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

Command Nightly

Model Overview

Cohere Command Nightly 实验版模型, 128K 上下文, 最新功能预览, 适合测试新特性。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Cohere nightly 2024-01-01 128K text text CC-BY-NC-4.0

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 85.2 % 5-shot
HumanEval 74.7 pass@1
GSM8K (Grade School Math 8K) 87.2 % 0-shot CoT
MATH 49.2 % 0-shot CoT
BBH (BIG-Bench Hard) 82.7 % 3-shot CoT
GPQA 45.5 % 0-shot
IFEval 80.8 % prompt_strict
ARC 95.9 % challenge
MUSR 57.5 % 0-shot
WinoGrande 80.8 % 0-shot

Pricing

Input Output Cache Read Cache Write

per million tokens

Strengths

  • MMLU score 85.2, strong knowledge reasoning.
  • GSM8K 87.2, robust math reasoning.
  • Enterprise-grade compliance.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging
  • Enterprise customer support

References