Claude Sonnet 3.6 (2022-10-22) vs GPT-4
Compare Claude Sonnet 3.6 (2022-10-22) by Anthropic against GPT-4 by OpenAI, context windows of 200K vs 8K, tested across 26 shared challenges.
Our Verdict
Claude Sonnet 3.6 (2022-10-22)
GPT-4
No community votes yet. On paper, these are closely matched - try both with your actual task to see which fits your workflow.
Claude Sonnet 3.6 (2022-10-22) is 4.0x cheaper per token — worth considering if cost matters.
Too close to call
API pricing
Cost per 1M tokens
Claude Sonnet 3.6 (2022-10-22)
Input
$3.00
10× cheaper
Output
$15.00
4.0× cheaper
GPT-4
Input
$30.00
Output
$60.00
Claude Sonnet 3.6 (2022-10-22) is cheaper on both — 10× input, 4.0× output
Writing DNA
Style Comparison
Similarity
52%
Claude Sonnet 3.6 (2022-10-22) uses 5.6x more headings
Claude Sonnet 3.6 (2022-10-22)
GPT-4
66%Vocabulary59%
74wSentence Length18w
0.72Hedging1.10
0.3Bold0.8
8.9Lists2.5
0.00Emoji0.00
0.06Headings0.00
0.09Transitions0.45
Based on 21 + 15 text responses
Research
What we learned reading every model
FAQ
