Compare · ModelsLive · 3 picked · head to head
Claude Fable 5.1 vs Claude Opus 5.5 vs GPT-6 Astra
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-6 Astra wins on 14/29 benchmarks
GPT-6 Astra wins 14 of 29 shared benchmarks. Leads in reasoning · math · knowledge.
Category leads
speed·Claude Opus 5.5agentic·Claude Opus 5.5reasoning·GPT-6 Astrageneral·Claude Opus 5.5coding·Claude Opus 5.5math·GPT-6 Astraknowledge·GPT-6 Astra
Hype vs Reality
Attention vs performance
Claude Fable 5.1
#16 by perf·#12 by attention
Claude Opus 5.5
#3 by perf·#9 by attention
GPT-6 Astra
#9 by perf·#1 by attention
Best value
Claude Opus 5.5
2.7x better value than GPT-6 Astra
Claude Fable 5.1
2.4 pts/$
$30.00/M
Claude Opus 5.5
6.9 pts/$
$12.00/M
GPT-6 Astra
2.6 pts/$
$30.00/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
Anthropic
$965.0B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
29 benchmarks · 3 models
Claude Fable 5.1Claude Opus 5.5GPT-6 Astra
Artificial Analysis · CritPt
Claude Fable 5.1
29.7
Claude Opus 5.5
31.7
GPT-6 Astra
31.7
Artificial Analysis · GDPval
Claude Opus 5.5 leads by +5.6
Claude Fable 5.1
61.7
Claude Opus 5.5
67.3
GPT-6 Astra
52.1
Artificial Analysis · Humanity's Last Exam
Claude Opus 5.5 leads by +2.3
Claude Fable 5.1
59.1
Claude Opus 5.5
61.4
GPT-6 Astra
54.7
Artificial Analysis · Long Context Reasoning
Claude Fable 5.1 leads by +0.6
Claude Fable 5.1
85.3
Claude Opus 5.5
84.7
GPT-6 Astra
80.7
Artificial Analysis · Quality Index
Claude Opus 5.5 leads by +4.3
Claude Fable 5.1
53.4
Claude Opus 5.5
57.6
GPT-6 Astra
52.7
Artificial Analysis · SciCode
Claude Opus 5.5 leads by +3.8
Claude Fable 5.1
63.1
Claude Opus 5.5
66.9
GPT-6 Astra
56.5
APEX-Agents
Claude Opus 5.5 leads by +4.9
APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.
Claude Fable 5.1
68.6
Claude Opus 5.5
73.5
GPT-6 Astra
64.7
ARC-AGI
GPT-6 Astra leads by +1.0
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Claude Fable 5.1
97.5
Claude Opus 5.5
97.5
GPT-6 Astra
98.5
ARC-AGI-2
GPT-6 Astra leads by +2.5
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Claude Fable 5.1
90.0
Claude Opus 5.5
92.5
GPT-6 Astra
95.0
Dtbench
Claude Opus 5.5 leads by +2.2
Claude Fable 5.1
96.0
Claude Opus 5.5
98.2
GPT-6 Astra
95.5
Ebr Bench
GPT-6 Astra leads by +4.8
Claude Fable 5.1
57.1
Claude Opus 5.5
71.4
GPT-6 Astra
76.2
Frontiercode
Claude Opus 5.5 leads by +1.4
Claude Fable 5.1
50.9
Claude Opus 5.5
54.6
GPT-6 Astra
53.3
FrontierMath-Tier-4-v2-Private
GPT-6 Astra leads by +2.6
Claude Fable 5.1
87.8
Claude Opus 5.5
95.0
GPT-6 Astra
97.6
FrontierMath-Tiers-1-3-v2-Private
GPT-6 Astra leads by +2.5
Claude Fable 5.1
90.2
Claude Opus 5.5
91.2
GPT-6 Astra
93.7
Frontierswe
GPT-6 Astra leads by +3.2
Claude Fable 5.1
56.3
Claude Opus 5.5
62.3
GPT-6 Astra
65.5
Furniture Assembly
Claude Opus 5.5 leads by +4.8
Claude Fable 5.1
57.1
Claude Opus 5.5
76.2
GPT-6 Astra
71.4
Lmca
Claude Opus 5.5 leads by +3.3
Claude Fable 5.1
77.0
Claude Opus 5.5
80.3
GPT-6 Astra
75.7
Mirrorcode
Claude Opus 5.5 leads by +4.1
Claude Fable 5.1
73.3
Claude Opus 5.5
77.4
GPT-6 Astra
46.7
Mystery Game Puzzles
GPT-6 Astra leads by +14.3
Claude Fable 5.1
53.7
Claude Opus 5.5
68.0
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 Fable 5.1
100.0
Claude Opus 5.5
100.0
GPT-6 Astra
100.0
Proofbench
Claude Fable 5.1
100.0
Claude Opus 5.5
100.0
GPT-6 Astra
99.0
SimpleQA Verified
GPT-6 Astra leads by +3.4
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Fable 5.1
70.8
Claude Opus 5.5
72.2
GPT-6 Astra
75.6
Artificial Analysis · GPQA Diamond
GPT-6 Astra leads by +2.4
Claude Fable 5.1
93.7
GPT-6 Astra
96.1
Artificial Analysis · MMMU Pro
Claude Opus 5.5 leads by +0.8
Claude Opus 5.5
87.7
GPT-6 Astra
86.9
Chess Puzzles
GPT-6 Astra leads by +26.3
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Fable 5.1
44.2
GPT-6 Astra
70.5
GPQA diamond
GPT-6 Astra leads by +6.9
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Opus 5.5
87.5
GPT-6 Astra
94.4
HLE
GPT-6 Astra leads by +8.7
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 Fable 5.1
43.8
GPT-6 Astra
52.5
Remote Labor Index
GPT-6 Astra leads by +2.9
Claude Fable 5.1
17.9
GPT-6 Astra
20.8
WeirdML
GPT-6 Astra leads by +0.7
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Claude Fable 5.1
92.9
GPT-6 Astra
93.6
Full benchmark table
| Benchmark | Claude Fable 5.1 | Claude Opus 5.5 | GPT-6 Astra |
|---|---|---|---|
Artificial Analysis · CritPt | 29.7 | 31.7 | 31.7 |
Artificial Analysis · GDPval | 61.7 | 67.3 | 52.1 |
Artificial Analysis · Humanity's Last Exam | 59.1 | 61.4 | 54.7 |
Artificial Analysis · Long Context Reasoning | 85.3 | 84.7 | 80.7 |
Artificial Analysis · Quality Index | 53.4 | 57.6 | 52.7 |
Artificial Analysis · SciCode | 63.1 | 66.9 | 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. | 68.6 | 73.5 | 64.7 |
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 97.5 | 97.5 | 98.5 |
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. | 90.0 | 92.5 | 95.0 |
Dtbench | 96.0 | 98.2 | 95.5 |
Ebr Bench | 57.1 | 71.4 | 76.2 |
Frontiercode | 50.9 | 54.6 | 53.3 |
FrontierMath-Tier-4-v2-Private | 87.8 | 95.0 | 97.6 |
FrontierMath-Tiers-1-3-v2-Private | 90.2 | 91.2 | 93.7 |
Frontierswe | 56.3 | 62.3 | 65.5 |
Furniture Assembly | 57.1 | 76.2 | 71.4 |
Lmca | 77.0 | 80.3 | 75.7 |
Mirrorcode | 73.3 | 77.4 | 46.7 |
Mystery Game Puzzles | 53.7 | 68.0 | 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 | 100.0 |
Proofbench | 100.0 | 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. | 70.8 | 72.2 | 75.6 |
Artificial Analysis · GPQA Diamond | 93.7 | — | 96.1 |
Artificial Analysis · MMMU Pro | — | 87.7 | 86.9 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 44.2 | — | 70.5 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | — | 87.5 | 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%. | 43.8 | — | 52.5 |
Remote Labor Index | 17.9 | — | 20.8 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 92.9 | — | 93.6 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $10.00 | $50.00 | 1.0M tokens (~500 books) | $200.00 | |
| $4.00 | $20.00 | 1.0M tokens (~500 books) | $80.00 | |
| $10.00 | $50.00 | 1.1M tokens (~525 books) | $200.00 |