You can now use Claude 3.5 Sonnet on Perplexity

Perplexity users can now select the Claude 3.5 Sonnet model, announced by Anthropic with high-speed AI task execution for new experiences.

The Sonnet is 2 times faster than Claude 3 Opus, its existing flagship model. Interestingly, the new Anthropic model brings more power than Claude 3 Opus and GPT-4o on the benchmark test. It shows that the model scored 8.9 from 9.0. OpenAI’s GPT-4o scored 8.77 points, Claude 3 Opus scored 8.61 points, and Sonnet 3 had 8.4 points.

Perplexity Benchmark Claude 3.5 Sonnet (Image Source – Perplexity)

Claude 3.5 Sonnet’s performance and reply speed wins over GPT-4o and Gemini 1.5 Pro models. The new LLM has topped different evaluations including:

• Graduate-level reasoning (GPQA)
• Undergraduate-level knowledge (MMLU)
• Coding proficiency (HumanEval)
• Multilingual math

It also outnumbered Claude 3 Opus (38%) by solving 64 percent of problems in agentic coding evaluation. Users with the Pro membership can check the new model while asking new questions or set the new model as the default. The new features will enable users to write or rewrite answer.

Perplexity AI Claude 3.5 Sonnet (Image Source – Perplexity)

Tap the rewrite button on the bottom left of the answer to select “Claude 3.5 Sonnet”. This step will be followed by an answer from Claude 3.5.

The next step is to set this new model as default, open Perplexity settings, and go to AI Model below the subscription option. Select the default AI model to Claude 3.5 Sonnet and all answers will be derived from the Anthropic’s model.

(source)