Compare · ModelsLive · 2 picked · head to head
Claude Sonnet 5.5 vs GPT-6 Astra
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-6 Astra wins on 12/20 benchmarks
GPT-6 Astra wins 12 of 20 shared benchmarks. Leads in knowledge · coding · math.
Category leads
speed·Claude Sonnet 5.5agentic·Claude Sonnet 5.5knowledge·GPT-6 Astracoding·GPT-6 Astramath·GPT-6 Astrageneral·GPT-6 Astra
Hype vs Reality
Attention vs performance
Claude Sonnet 5.5
#34 by perf·#12 by attention
GPT-6 Astra
#10 by perf·#5 by attention
Best value
Claude Sonnet 5.5
4.4x better value than GPT-6 Astra
Claude Sonnet 5.5
11.0 pts/$
$6.00/M
GPT-6 Astra
2.5 pts/$
$30.00/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
20 benchmarks · 2 models
Claude Sonnet 5.5GPT-6 Astra
Artificial Analysis · CritPt
GPT-6 Astra leads by +0.3
Claude Sonnet 5.5
31.4
GPT-6 Astra
31.7
Artificial Analysis · GDPval
Claude Sonnet 5.5 leads by +13.3
Claude Sonnet 5.5
66.9
GPT-6 Astra
53.6
Artificial Analysis · Humanity's Last Exam
Claude Sonnet 5.5 leads by +0.3
Claude Sonnet 5.5
55.0
GPT-6 Astra
54.7
Artificial Analysis · Long Context Reasoning
Claude Sonnet 5.5 leads by +2.0
Claude Sonnet 5.5
82.7
GPT-6 Astra
80.7
Artificial Analysis · Quality Index
Claude Sonnet 5.5 leads by +3.3
Claude Sonnet 5.5
56.0
GPT-6 Astra
52.7
Artificial Analysis · SciCode
Claude Sonnet 5.5 leads by +4.5
Claude Sonnet 5.5
61.0
GPT-6 Astra
56.5
APEX-Agents
Claude Sonnet 5.5 leads by +10.8
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Claude Sonnet 5.5
75.5
GPT-6 Astra
64.7
Chess Puzzles
GPT-6 Astra leads by +15.8
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-6 Astra
72.6
Frontiercode
GPT-6 Astra leads by +1.2
Claude Sonnet 5.5
52.1
GPT-6 Astra
53.3
FrontierMath-Tier-4-v2-Private
GPT-6 Astra leads by +17.1
Claude Sonnet 5.5
80.5
GPT-6 Astra
97.6
FrontierMath-Tiers-1-3-v2-Private
GPT-6 Astra leads by +4.9
Claude Sonnet 5.5
88.8
GPT-6 Astra
93.7
Frontierswe
GPT-6 Astra leads by +3.6
Claude Sonnet 5.5
61.9
GPT-6 Astra
65.5
Furniture Assembly
GPT-6 Astra leads by +7.1
Claude Sonnet 5.5
64.3
GPT-6 Astra
71.4
GPQA diamond
GPT-6 Astra leads by +0.3
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-6 Astra
94.4
HLE
GPT-6 Astra leads by +9.1
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-6 Astra
52.5
Mystery Game Puzzles
GPT-6 Astra leads by +20.9
Claude Sonnet 5.5
61.5
GPT-6 Astra
82.4
OTIS Mock AIME 2024-2025
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Sonnet 5.5
100.0
GPT-6 Astra
100.0
Proofbench
Claude Sonnet 5.5 leads by +1.0
Claude Sonnet 5.5
100.0
GPT-6 Astra
99.0
SimpleQA Verified
GPT-6 Astra leads by +29.1
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-6 Astra
75.6
SpireClimb
GPT-6 Astra leads by +24.3
Claude Sonnet 5.5
1.3
GPT-6 Astra
25.7
Full benchmark table
| Benchmark | Claude Sonnet 5.5 | GPT-6 Astra |
|---|---|---|
Artificial Analysis · CritPt | 31.4 | 31.7 |
Artificial Analysis · GDPval | 66.9 | 53.6 |
Artificial Analysis · Humanity's Last Exam | 55.0 | 54.7 |
Artificial Analysis · Long Context Reasoning | 82.7 | 80.7 |
Artificial Analysis · Quality Index | 56.0 | 52.7 |
Artificial Analysis · SciCode | 61.0 | 56.5 |
APEX-Agents APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments. | 75.5 | 64.7 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 56.9 | 72.6 |
Frontiercode | 52.1 | 53.3 |
FrontierMath-Tier-4-v2-Private | 80.5 | 97.6 |
FrontierMath-Tiers-1-3-v2-Private | 88.8 | 93.7 |
Frontierswe | 61.9 | 65.5 |
Furniture Assembly | 64.3 | 71.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 94.1 | 94.4 |
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 | 52.5 |
Mystery Game Puzzles | 61.5 | 82.4 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 100.0 | 100.0 |
Proofbench | 100.0 | 99.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 46.5 | 75.6 |
SpireClimb | 1.3 | 25.7 |
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 | |
| $10.00 | $50.00 | 1.1M tokens (~525 books) | $200.00 |