Source URL: https://simonwillison.net/2025/May/22/updated-anthropic-models/#atom-everything
Source: Simon Willison’s Weblog
Title: Updated Anthropic model comparison table
Feedly Summary: Updated Anthropic model comparison table
A few details in here about Claude 4 that I hadn’t spotted elsewhere:
The training cut-off date for Claude Opus 4 and Claude Sonnet 4 is March 2025! That’s the most recent cut-off for any of the current popular models, really impressive.
Opus 4 has a max output of 32,000 tokens, Sonnet 4 has a max output of 64,000 tokens. Claude 3.7 Sonnet is 64,000 tokens too, so this is a small regression for Opus.
The input limit for both of the Claude 4 models is still stuck at 200,000. I’m disjointed by this, I was hoping for a leap to a million to catch up with GPT 4.1 and the Gemini Pro series.
Claude 3 Haiku is still in that table – it remains Anthropic’s cheapest model, priced slightly lower than Claude 3.5 Haiku.
For pricing: Sonnet 4 is the same price as Sonnet 3.7 ($3/million input, $15/million output). Opus 4 matches the pricing of the older Opus 3 – $15/million for input and $75/million for output. I’ve updated llm-prices.com with the new models.
Tags: anthropic, claude, generative-ai, ai, llms, llm-pricing
AI Summary and Description: Yes
Summary: The text provides insights into the latest Claude models from Anthropic, particularly highlighting performance metrics, pricing, and token limits. These details are crucial for professionals focusing on generative AI and LLMs in terms of evaluating model capabilities and market competitiveness.
Detailed Description: The provided text compares recent developments in the Claude models by Anthropic, specifically Claude Opus 4 and Claude Sonnet 4. Here are the key points outlined:
– **Training Cut-Off Date**:
– Claude Opus 4 and Claude Sonnet 4 have a training cut-off in March 2025, indicating these models have access to more recent data compared to competitors.
– **Token Limits**:
– Opus 4 supports a maximum output of 32,000 tokens.
– Sonnet 4 allows a max output of 64,000 tokens, consistent with Claude 3.7 Sonnet, although Opus has seen a regression in token output.
– Both models retain an input limit of 200,000 tokens, which the author finds disappointing, particularly in comparison to advancements seen with other models like GPT 4.1 and the Gemini Pro series.
– **Cost Considerations**:
– Claude 3.7 Haiku remains the most affordable option presented in the comparison table.
– Pricing for Sonnet 4 is unchanged from Sonnet 3.7, established at $3/million for input and $15/million for output.
– Opus 4 matches the older Opus 3 pricing model of $15/million for input and $75/million for output.
– **Market Update**:
– The author mentions updating llm-prices.com with the new models, reflecting ongoing competition in the LLM landscape.
These insights are particularly valuable for security and compliance professionals working with AI, as they need to consider not just the technical capabilities of models but also the implications of pricing and data usage limits on their operational workflows and compliance with regulations governing the deployment of AI technologies.