LLM Reference

Claude 2

About

Claude 2, crafted by Anthropic, represents a leap forward in large language model technology, striving to be helpful, harmless, and honest. It's recognized for its improved performance, longer responses, and a remarkable 100,000-token context window, allowing it to handle extensive text inputs with ease. Excelling in coding, math, and reasoning tasks, Claude 2 performs exceptionally on standardized tests like the Bar exam and GRE. Despite its achievements, it faces challenges such as generating inaccurate information ("hallucinations") and simulating emotional responses rather than possessing true empathy. The model's knowledge extends up to early 2023, so it might miss the latest updates. While not immune to "jailbreaks," it aims to enhance AI safety significantly 135.

Capabilities

MultimodalFunction CallingTool UseJSON Mode

Providers(2)

ProviderInput (per 1M)Output (per 1M)Type
AWS Bedrock$8$24
Serverless
Anthropic Console$8$24
Serverless

Specifications

FamilyClaude 2
Released2023-07-11
Context100K
ArchitectureDecoder Only
Knowledge cutoff2023-01
Specializationgeneral