Compare · ModelsLive · 2 picked · head to head
Claude Sonnet 5.5 vs GPT-5.4 Pro
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Sonnet 5.5 wins on 6/6 benchmarks
Claude Sonnet 5.5 wins 6 of 6 shared benchmarks. Leads in knowledge · math.
Category leads
knowledge·Claude Sonnet 5.5math·Claude Sonnet 5.5
Hype vs Reality
Attention vs performance
Claude Sonnet 5.5
#34 by perf·#12 by attention
GPT-5.4 Pro
#40 by perf·#10 by attention
Best value
Claude Sonnet 5.5
18.0x better value than GPT-5.4 Pro
Claude Sonnet 5.5
11.0 pts/$
$6.00/M
GPT-5.4 Pro
0.6 pts/$
$105.00/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
6 benchmarks · 2 models
Claude Sonnet 5.5GPT-5.4 Pro
Chess Puzzles
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Sonnet 5.5
56.9
GPT-5.4 Pro
56.9
FrontierMath-Tier-4-v2-Private
Claude Sonnet 5.5 leads by +21.9
Claude Sonnet 5.5
80.5
GPT-5.4 Pro
58.5
FrontierMath-Tiers-1-3-v2-Private
Claude Sonnet 5.5 leads by +6.3
Claude Sonnet 5.5
88.8
GPT-5.4 Pro
82.5
GPQA diamond
Claude Sonnet 5.5 leads by +1.3
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Sonnet 5.5
94.1
GPT-5.4 Pro
92.8
HLE
Claude Sonnet 5.5 leads by +1.9
HLE (Humanity's Last Exam) · a reasoning benchmark designed to be the hardest public evaluation of AI. Questions span mathematics, physics, philosophy, and logic · curated to be at or beyond the frontier of human expert capability. Tested with and without tool augmentation. Claude Opus 4.7 scores 46.9% without tools and 54.7% with tools · making it one of the few benchmarks where the top score is below 60%.
Claude Sonnet 5.5
43.4
GPT-5.4 Pro
41.5
SimpleQA Verified
Claude Sonnet 5.5 leads by +0.2
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Sonnet 5.5
46.5
GPT-5.4 Pro
46.3
Full benchmark table
| Benchmark | Claude Sonnet 5.5 | GPT-5.4 Pro |
|---|---|---|
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 56.9 | 56.9 |
FrontierMath-Tier-4-v2-Private | 80.5 | 58.5 |
FrontierMath-Tiers-1-3-v2-Private | 88.8 | 82.5 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 94.1 | 92.8 |
HLE HLE (Humanity's Last Exam) · a reasoning benchmark designed to be the hardest public evaluation of AI. Questions span mathematics, physics, philosophy, and logic · curated to be at or beyond the frontier of human expert capability. Tested with and without tool augmentation. Claude Opus 4.7 scores 46.9% without tools and 54.7% with tools · making it one of the few benchmarks where the top score is below 60%. | 43.4 | 41.5 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 46.5 | 46.3 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $2.00 | $10.00 | 1.0M tokens (~500 books) | $40.00 | |
| $30.00 | $180.00 | 1.1M tokens (~525 books) | $675.00 |
People also compared
Gemini 3.1 Pro Preview vs GPT-5.4 ProClaude Sonnet 5.5 vs phi-3-small 7.4BClaude Sonnet 5.5 vs GPT-5.6 SolClaude Sonnet 5.5 vs GPT-5.6 TerraClaude Sonnet 5.5 vs gpt-oss-20b (free)Claude Sonnet 5.5 vs Qwen3 30B A3B Thinking 2507Claude Sonnet 5.5 vs GPT-4 Turbo (older v1106)Claude Sonnet 5.5 vs Grok 3 Mini Beta