Compare · ModelsLive · 2 picked · head to head
Claude Opus 5 vs GPT-6 Sol
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Claude Opus 5 wins on 9/16 benchmarks
Claude Opus 5 wins 9 of 16 shared benchmarks. Leads in agentic · reasoning · general.
Category leads
agentic·Claude Opus 5reasoning·Claude Opus 5knowledge·GPT-6 Solgeneral·Claude Opus 5coding·Claude Opus 5math·GPT-6 Sol
Hype vs Reality
Attention vs performance
Claude Opus 5
#21 by perf·#9 by attention
GPT-6 Sol
#23 by perf·#5 by attention
Best value
GPT-6 Sol
2.5x better value than Claude Opus 5
Claude Opus 5
4.6 pts/$
$15.00/M
GPT-6 Sol
11.5 pts/$
$6.00/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
16 benchmarks · 2 models
Claude Opus 5GPT-6 Sol
APEX-Agents
Claude Opus 5 leads by +11.5
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Claude Opus 5
65.8
GPT-6 Sol
54.3
ARC-AGI
Claude Opus 5 leads by +2.0
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude Opus 5
97.5
GPT-6 Sol
95.5
ARC-AGI-2
Claude Opus 5 leads by +0.8
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Claude Opus 5
90.4
GPT-6 Sol
89.6
Chess Puzzles
GPT-6 Sol leads by +13.7
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Opus 5
40.0
GPT-6 Sol
53.7
Dtbench
Claude Opus 5 leads by +0.9
Claude Opus 5
96.5
GPT-6 Sol
95.5
Ebr Bench
GPT-6 Sol leads by +7.6
Claude Opus 5
45.7
GPT-6 Sol
53.3
Frontiercode
Claude Opus 5 leads by +4.1
Claude Opus 5
53.4
GPT-6 Sol
49.3
FrontierMath-Tier-4-v2-Private
GPT-6 Sol leads by +16.8
Claude Opus 5
73.2
GPT-6 Sol
90.0
FrontierMath-Tiers-1-3-v2-Private
GPT-6 Sol leads by +4.2
Claude Opus 5
85.6
GPT-6 Sol
89.8
Furniture Assembly
Claude Opus 5 leads by +3.6
Claude Opus 5
44.0
GPT-6 Sol
40.5
GPQA diamond
GPT-6 Sol leads by +0.6
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Opus 5
91.8
GPT-6 Sol
92.4
Lmca
Claude Opus 5 leads by +6.3
Claude Opus 5
75.8
GPT-6 Sol
69.5
Mystery Game Puzzles
Claude Opus 5 leads by +3.3
Claude Opus 5
54.8
GPT-6 Sol
51.5
OTIS Mock AIME 2024-2025
GPT-6 Sol leads by +1.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Opus 5
98.9
GPT-6 Sol
100.0
Proofbench
Claude Opus 5 leads by +16.0
Claude Opus 5
99.0
GPT-6 Sol
83.0
SimpleQA Verified
GPT-6 Sol leads by +0.8
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Opus 5
59.9
GPT-6 Sol
60.7
Full benchmark table
| Benchmark | Claude Opus 5 | GPT-6 Sol |
|---|---|---|
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 65.8 | 54.3 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 97.5 | 95.5 |
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. | 90.4 | 89.6 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 40.0 | 53.7 |
Dtbench | 96.5 | 95.5 |
Ebr Bench | 45.7 | 53.3 |
Frontiercode | 53.4 | 49.3 |
FrontierMath-Tier-4-v2-Private | 73.2 | 90.0 |
FrontierMath-Tiers-1-3-v2-Private | 85.6 | 89.8 |
Furniture Assembly | 44.0 | 40.5 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 91.8 | 92.4 |
Lmca | 75.8 | 69.5 |
Mystery Game Puzzles | 54.8 | 51.5 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 98.9 | 100.0 |
Proofbench | 99.0 | 83.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 59.9 | 60.7 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $5.00 | $25.00 | 1.0M tokens (~500 books) | $100.00 | |
| $2.00 | $10.00 | 1.1M tokens (~525 books) | $40.00 |