Compare · ModelsLive · 2 picked · head to head
Claude Sonnet 5 vs GPT-5.4
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-5.4 wins on 10/14 benchmarks
GPT-5.4 wins 10 of 14 shared benchmarks. Leads in knowledge · general · math.
Category leads
agentic·Claude Sonnet 5knowledge·GPT-5.4coding·Claude Sonnet 5general·GPT-5.4math·GPT-5.4
Hype vs Reality
Attention vs performance
Claude Sonnet 5
#66 by perf·#10 by attention
GPT-5.4
#104 by perf·#4 by attention
Best value
Claude Sonnet 5
1.6x better value than GPT-5.4
Claude Sonnet 5
9.5 pts/$
$6.00/M
GPT-5.4
6.0 pts/$
$8.75/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
14 benchmarks · 2 models
Claude Sonnet 5GPT-5.4
APEX-Agents
Claude Sonnet 5 leads by +2.1
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Claude Sonnet 5
54.5
GPT-5.4
52.4
Chess Puzzles
GPT-5.4 leads by +9.5
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Sonnet 5
31.6
GPT-5.4
41.1
Deepswe
Claude Sonnet 5 leads by +2.1
Claude Sonnet 5
53.9
GPT-5.4
51.8
Dtbench
GPT-5.4 leads by +3.2
Claude Sonnet 5
87.5
GPT-5.4
90.7
FrontierMath-Tier-4-v2-Private
GPT-5.4 leads by +19.7
Claude Sonnet 5
29.3
GPT-5.4
49.0
FrontierMath-Tiers-1-3-v2-Private
GPT-5.4 leads by +13.0
Claude Sonnet 5
65.6
GPT-5.4
78.6
GPQA diamond
GPT-5.4 leads by +3.7
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
87.4
GPT-5.4
91.1
GSO-Bench
Claude Sonnet 5 leads by +5.9
GSO-Bench · evaluates AI models on real-world open-source software engineering tasks, testing the ability to understand and resolve actual GitHub issues.
Claude Sonnet 5
37.3
GPT-5.4
31.4
Lmca
GPT-5.4 leads by +2.3
Claude Sonnet 5
58.9
GPT-5.4
61.1
Mystery Game Puzzles
GPT-5.4 leads by +2.2
Claude Sonnet 5
28.4
GPT-5.4
30.6
OTIS Mock AIME 2024-2025
GPT-5.4 leads by +3.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Sonnet 5
94.7
GPT-5.4
97.8
Proofbench
Claude Sonnet 5 leads by +21.0
Claude Sonnet 5
77.0
GPT-5.4
56.0
SimpleQA Verified
GPT-5.4 leads by +11.4
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Sonnet 5
33.7
GPT-5.4
45.1
WeirdML
GPT-5.4 leads by +8.9
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Claude Sonnet 5
68.8
GPT-5.4
77.7
Full benchmark table
| Benchmark | Claude Sonnet 5 | GPT-5.4 |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 54.5 | 52.4 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 31.6 | 41.1 |
Deepswe | 53.9 | 51.8 |
Dtbench | 87.5 | 90.7 |
FrontierMath-Tier-4-v2-Private | 29.3 | 49.0 |
FrontierMath-Tiers-1-3-v2-Private | 65.6 | 78.6 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 87.4 | 91.1 |
GSO-Bench GSO-Bench · evaluates AI models on real-world open-source software engineering tasks, testing the ability to understand and resolve actual GitHub issues. | 37.3 | 31.4 |
Lmca | 58.9 | 61.1 |
Mystery Game Puzzles | 28.4 | 30.6 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 94.7 | 97.8 |
Proofbench | 77.0 | 56.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 33.7 | 45.1 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 68.8 | 77.7 |
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 | |
| $2.50 | $15.00 | 1.1M tokens (~525 books) | $56.25 |