Compare · ModelsLive · 2 picked · head to head
Claude 3.7 Sonnet vs Claude Sonnet 4.6
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Sonnet 4.6 wins on 6/6 benchmarks
Claude Sonnet 4.6 wins 6 of 6 shared benchmarks. Leads in reasoning · math · knowledge.
Category leads
reasoning·Claude Sonnet 4.6math·Claude Sonnet 4.6knowledge·Claude Sonnet 4.6coding·Claude Sonnet 4.6
Hype vs Reality
Attention vs performance
Claude 3.7 Sonnet
#103 by perf·no signal
Claude Sonnet 4.6
#104 by perf·#18 by attention
Best value
Claude 3.7 Sonnet
1.0x better value than Claude Sonnet 4.6
Claude 3.7 Sonnet
5.3 pts/$
$9.00/M
Claude Sonnet 4.6
5.3 pts/$
$9.00/M
Vendor risk
Who is behind the model
Anthropic
$380.0B·Tier 1
Anthropic
$380.0B·Tier 1
Head to head
6 benchmarks · 2 models
Claude 3.7 SonnetClaude Sonnet 4.6
ARC-AGI
Claude Sonnet 4.6 leads by +57.9
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude 3.7 Sonnet
28.6
Claude Sonnet 4.6
86.5
ARC-AGI-2
Claude Sonnet 4.6 leads by +59.5
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Claude 3.7 Sonnet
0.9
Claude Sonnet 4.6
60.4
FrontierMath-2025-02-28-Private
Claude Sonnet 4.6 leads by +28.3
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Claude 3.7 Sonnet
4.1
Claude Sonnet 4.6
32.4
GPQA diamond
Claude Sonnet 4.6 leads by +10.2
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude 3.7 Sonnet
73.0
Claude Sonnet 4.6
83.2
OTIS Mock AIME 2024-2025
Claude Sonnet 4.6 leads by +28.0
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude 3.7 Sonnet
57.7
Claude Sonnet 4.6
85.8
SWE-Bench verified
Claude Sonnet 4.6 leads by +14.3
SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability.
Claude 3.7 Sonnet
61.0
Claude Sonnet 4.6
75.2
Full benchmark table
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 28.6 | 86.5 |
ARC-AGI-2 ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data. | 0.9 | 60.4 |
FrontierMath-2025-02-28-Private FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning. | 4.1 | 32.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 73.0 | 83.2 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 57.7 | 85.8 |
SWE-Bench verified SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability. | 61.0 | 75.2 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $3.00 | $15.00 | 200K tokens (~100 books) | $60.00 | |
| $3.00 | $15.00 | 1.0M tokens (~500 books) | $60.00 |