Compare · ModelsLive · 3 picked · head to head
GPT-5.4 Pro vs GPT-5.4 vs Claude Opus 4.6
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-5.4 Pro wins on 8/14 benchmarks
GPT-5.4 Pro wins 8 of 14 shared benchmarks. Leads in reasoning · knowledge · math.
Category leads
reasoning·GPT-5.4 Proknowledge·GPT-5.4 Promath·GPT-5.4 Proagentic·GPT-5.4arena·Claude Opus 4.6coding·Claude Opus 4.6
Hype vs Reality
Attention vs performance
GPT-5.4 Pro
#26 by perf·no signal
GPT-5.4
#46 by perf·no signal
Claude Opus 4.6
#56 by perf·#4 by attention
Best value
GPT-5.4
1.8x better value than Claude Opus 4.6
GPT-5.4 Pro
0.6 pts/$
$105.00/M
GPT-5.4
6.7 pts/$
$8.75/M
Claude Opus 4.6
3.8 pts/$
$15.00/M
Vendor risk
Who is behind the model
OpenAI
$840.0B·Tier 1
OpenAI
$840.0B·Tier 1
Anthropic
$380.0B·Tier 1
Head to head
14 benchmarks · 3 models
GPT-5.4 ProGPT-5.4Claude Opus 4.6
ARC-AGI
GPT-5.4 Pro leads by +0.5
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
GPT-5.4 Pro
94.5
GPT-5.4
93.7
Claude Opus 4.6
94.0
ARC-AGI-2
GPT-5.4 Pro leads by +9.4
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
GPT-5.4 Pro
83.3
GPT-5.4
74.0
Claude Opus 4.6
69.2
Chess Puzzles
GPT-5.4 Pro leads by +14.6
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
GPT-5.4 Pro
58.6
GPT-5.4
44.0
Claude Opus 4.6
17.0
FrontierMath-2025-02-28-Private
GPT-5.4 Pro leads by +2.4
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
GPT-5.4 Pro
50.0
GPT-5.4
47.6
Claude Opus 4.6
40.7
FrontierMath-Tier-4-2025-07-01-Private
GPT-5.4 Pro leads by +10.4
FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning.
GPT-5.4 Pro
37.5
GPT-5.4
27.1
Claude Opus 4.6
22.9
GPQA diamond
GPT-5.4 Pro leads by +1.7
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
GPT-5.4 Pro
92.8
GPT-5.4
91.1
Claude Opus 4.6
87.4
SimpleQA Verified
GPT-5.4 Pro leads by +1.3
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
GPT-5.4 Pro
47.8
GPT-5.4
44.8
Claude Opus 4.6
46.5
APEX-Agents
GPT-5.4 leads by +4.2
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
GPT-5.4
35.9
Claude Opus 4.6
31.7
Chatbot Arena Elo · Overall
Claude Opus 4.6 leads by +30.8
GPT-5.4
1465.8
Claude Opus 4.6
1496.6
OTIS Mock AIME 2024-2025
GPT-5.4 leads by +0.9
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
GPT-5.4
95.3
Claude Opus 4.6
94.4
PostTrainBench
Claude Opus 4.6 leads by +2.9
GPT-5.4
20.2
Claude Opus 4.6
23.2
SimpleBench
GPT-5.4 Pro leads by +7.8
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
GPT-5.4 Pro
68.9
Claude Opus 4.6
61.1
SWE-Bench verified
Claude Opus 4.6 leads by +1.9
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.
GPT-5.4
76.9
Claude Opus 4.6
78.7
WeirdML
Claude Opus 4.6 leads by +20.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
GPT-5.4
57.4
Claude Opus 4.6
77.9
Full benchmark table
| Benchmark | GPT-5.4 Pro | GPT-5.4 | Claude Opus 4.6 |
|---|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 94.5 | 93.7 | 94.0 |
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. | 83.3 | 74.0 | 69.2 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 58.6 | 44.0 | 17.0 |
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. | 50.0 | 47.6 | 40.7 |
FrontierMath-Tier-4-2025-07-01-Private FrontierMath Tier 4 (Jul 2025) · the most challenging tier of frontier mathematics, containing problems that push the absolute limits of AI mathematical reasoning. | 37.5 | 27.1 | 22.9 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 92.8 | 91.1 | 87.4 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 47.8 | 44.8 | 46.5 |
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | — | 35.9 | 31.7 |
Chatbot Arena Elo · Overall | — | 1465.8 | 1496.6 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | — | 95.3 | 94.4 |
PostTrainBench | — | 20.2 | 23.2 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 68.9 | — | 61.1 |
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. | — | 76.9 | 78.7 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | — | 57.4 | 77.9 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $30.00 | $180.00 | 1.1M tokens (~525 books) | $675.00 | |
| $2.50 | $15.00 | 1.1M tokens (~525 books) | $56.25 | |
| $5.00 | $25.00 | 1.0M tokens (~500 books) | $100.00 |