Compare · ModelsLive · 3 picked · head to head
Claude Fable 5.1 vs Gemini 3.1 Pro Preview vs GPT-6 Astra
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
GPT-6 Astra wins on 22/37 benchmarks
GPT-6 Astra wins 22 of 37 shared benchmarks. Leads in reasoning · knowledge · general.
Category leads
speed·Claude Fable 5.1agentic·Claude Fable 5.1reasoning·GPT-6 Astraknowledge·GPT-6 Astrageneral·GPT-6 Astramath·GPT-6 Astracoding·GPT-6 Astra
Hype vs Reality
Attention vs performance
Claude Fable 5.1
#23 by perf·#5 by attention
Gemini 3.1 Pro Preview
#143 by perf·#10 by attention
GPT-6 Astra
#9 by perf·#2 by attention
Best value
Gemini 3.1 Pro Preview
2.8x better value than GPT-6 Astra
Claude Fable 5.1
2.3 pts/$
$30.00/M
Gemini 3.1 Pro Preview
7.0 pts/$
$7.00/M
GPT-6 Astra
2.5 pts/$
$30.00/M
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
Google DeepMind
$3.61T·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
37 benchmarks · 3 models
Claude Fable 5.1Gemini 3.1 Pro PreviewGPT-6 Astra
Artificial Analysis · CritPt
GPT-6 Astra leads by +2.0
Claude Fable 5.1
29.7
Gemini 3.1 Pro Preview
17.7
GPT-6 Astra
31.7
Artificial Analysis · GDPval
Claude Fable 5.1 leads by +9.1
Claude Fable 5.1
62.7
Gemini 3.1 Pro Preview
14.6
GPT-6 Astra
53.6
Artificial Analysis · GPQA Diamond
GPT-6 Astra leads by +2.0
Claude Fable 5.1
93.7
Gemini 3.1 Pro Preview
94.1
GPT-6 Astra
96.1
Artificial Analysis · Humanity's Last Exam
Claude Fable 5.1 leads by +4.4
Claude Fable 5.1
59.1
Gemini 3.1 Pro Preview
47.0
GPT-6 Astra
54.7
Artificial Analysis · Long Context Reasoning
Claude Fable 5.1 leads by +3.3
Claude Fable 5.1
85.3
Gemini 3.1 Pro Preview
82.0
GPT-6 Astra
80.7
Artificial Analysis · Quality Index
Claude Fable 5.1 leads by +0.7
Claude Fable 5.1
53.4
Gemini 3.1 Pro Preview
29.7
GPT-6 Astra
52.7
Artificial Analysis · SciCode
Claude Fable 5.1 leads by +4.4
Claude Fable 5.1
63.1
Gemini 3.1 Pro Preview
58.7
GPT-6 Astra
56.5
APEX-Agents
Claude Fable 5.1 leads by +3.9
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
Gemini 3.1 Pro Preview
35.1
GPT-6 Astra
64.7
ARC-AGI
GPT-6 Astra leads by +0.5
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
Gemini 3.1 Pro Preview
98.0
GPT-6 Astra
98.5
ARC-AGI-2
GPT-6 Astra leads by +5.0
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
Gemini 3.1 Pro Preview
77.1
GPT-6 Astra
95.0
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
45.3
Gemini 3.1 Pro Preview
48.4
GPT-6 Astra
74.8
Dtbench
Claude Fable 5.1 leads by +0.5
Claude Fable 5.1
96.0
Gemini 3.1 Pro Preview
95.1
GPT-6 Astra
95.5
Ebr Bench
GPT-6 Astra leads by +19.0
Claude Fable 5.1
57.1
Gemini 3.1 Pro Preview
14.3
GPT-6 Astra
76.2
FrontierMath-Tier-4-v2-Private
GPT-6 Astra leads by +9.8
Claude Fable 5.1
87.8
Gemini 3.1 Pro Preview
26.8
GPT-6 Astra
97.6
FrontierMath-Tiers-1-3-v2-Private
GPT-6 Astra leads by +3.7
Claude Fable 5.1
94.9
Gemini 3.1 Pro Preview
62.8
GPT-6 Astra
98.6
Furniture Assembly
GPT-6 Astra leads by +14.3
Claude Fable 5.1
57.1
Gemini 3.1 Pro Preview
0.0
GPT-6 Astra
71.4
GSO-Bench
Claude Fable 5.1 leads by +8.8
GSO-Bench · evaluates AI models on real-world open-source software engineering tasks, testing the ability to understand and resolve actual GitHub issues.
Claude Fable 5.1
88.2
Gemini 3.1 Pro Preview
22.6
GPT-6 Astra
79.4
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 Fable 5.1
43.8
Gemini 3.1 Pro Preview
43.7
GPT-6 Astra
52.5
Lmca
Claude Fable 5.1 leads by +1.3
Claude Fable 5.1
77.0
Gemini 3.1 Pro Preview
63.3
GPT-6 Astra
75.7
Mirrorcode
Claude Fable 5.1 leads by +33.3
Claude Fable 5.1
93.3
Gemini 3.1 Pro Preview
8.9
GPT-6 Astra
60.0
Mystery Game Puzzles
GPT-6 Astra leads by +28.6
Claude Fable 5.1
53.7
Gemini 3.1 Pro Preview
27.3
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
Gemini 3.1 Pro Preview
95.6
GPT-6 Astra
100.0
PostTrainBench
Claude Fable 5.1 leads by +2.8
Claude Fable 5.1
44.6
Gemini 3.1 Pro Preview
18.1
GPT-6 Astra
41.9
Proofbench
Claude Fable 5.1 leads by +1.0
Claude Fable 5.1
100.0
Gemini 3.1 Pro Preview
26.0
GPT-6 Astra
99.0
SimpleBench
Claude Fable 5.1 leads by +0.1
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Claude Fable 5.1
83.9
Gemini 3.1 Pro Preview
75.5
GPT-6 Astra
83.8
SimpleQA Verified
GPT-6 Astra leads by +2.1
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
Gemini 3.1 Pro Preview
73.5
GPT-6 Astra
75.6
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
Gemini 3.1 Pro Preview
72.1
GPT-6 Astra
93.6
Artificial Analysis · MMMU Pro
GPT-6 Astra leads by +4.5
Gemini 3.1 Pro Preview
82.4
GPT-6 Astra
86.9
Balrog
GPT-6 Astra leads by +11.3
Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning.
Gemini 3.1 Pro Preview
57.0
GPT-6 Astra
68.3
Deepswe
GPT-6 Astra leads by +62.4
Gemini 3.1 Pro Preview
11.7
GPT-6 Astra
74.1
Frontiercode
GPT-6 Astra leads by +2.4
Claude Fable 5.1
50.9
GPT-6 Astra
53.3
Frontierswe
GPT-6 Astra leads by +9.2
Claude Fable 5.1
56.3
GPT-6 Astra
65.5
GeoBench
GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding.
Claude Fable 5.1
84.0
Gemini 3.1 Pro Preview
84.0
GPQA diamond
GPT-6 Astra leads by +1.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Gemini 3.1 Pro Preview
96.5
GPT-6 Astra
98.3
Remote Labor Index
GPT-6 Astra leads by +2.9
Claude Fable 5.1
17.9
GPT-6 Astra
20.8
SpireClimb
GPT-6 Astra leads by +32.0
Claude Fable 5.1
2.0
GPT-6 Astra
34.0
WeirdML v3
GPT-6 Astra leads by +11.7
Claude Fable 5.1
48.4
GPT-6 Astra
60.0
Full benchmark table
| Benchmark | Claude Fable 5.1 | Gemini 3.1 Pro Preview | GPT-6 Astra |
|---|---|---|---|
Artificial Analysis · CritPt | 29.7 | 17.7 | 31.7 |
Artificial Analysis · GDPval | 62.7 | 14.6 | 53.6 |
Artificial Analysis · GPQA Diamond | 93.7 | 94.1 | 96.1 |
Artificial Analysis · Humanity's Last Exam | 59.1 | 47.0 | 54.7 |
Artificial Analysis · Long Context Reasoning | 85.3 | 82.0 | 80.7 |
Artificial Analysis · Quality Index | 53.4 | 29.7 | 52.7 |
Artificial Analysis · SciCode | 63.1 | 58.7 | 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 | 35.1 | 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 | 98.0 | 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 | 77.1 | 95.0 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 45.3 | 48.4 | 74.8 |
Dtbench | 96.0 | 95.1 | 95.5 |
Ebr Bench | 57.1 | 14.3 | 76.2 |
FrontierMath-Tier-4-v2-Private | 87.8 | 26.8 | 97.6 |
FrontierMath-Tiers-1-3-v2-Private | 94.9 | 62.8 | 98.6 |
Furniture Assembly | 57.1 | 0.0 | 71.4 |
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. | 88.2 | 22.6 | 79.4 |
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. | 43.8 | 43.7 | 52.5 |
Lmca | 77.0 | 63.3 | 75.7 |
Mirrorcode | 93.3 | 8.9 | 60.0 |
Mystery Game Puzzles | 53.7 | 27.3 | 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 | 95.6 | 100.0 |
PostTrainBench | 44.6 | 18.1 | 41.9 |
Proofbench | 100.0 | 26.0 | 99.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 83.9 | 75.5 | 83.8 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 70.8 | 73.5 | 75.6 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 92.9 | 72.1 | 93.6 |
Artificial Analysis · MMMU Pro | n/a | 82.4 | 86.9 |
Balrog Balrog · benchmarks AI agents on text-based adventure games, testing language understanding, strategic planning, and long-horizon reasoning. | n/a | 57.0 | 68.3 |
Deepswe | n/a | 11.7 | 74.1 |
Frontiercode | 50.9 | n/a | 53.3 |
Frontierswe | 56.3 | n/a | 65.5 |
GeoBench GeoBench · tests geographic knowledge and spatial reasoning across countries, landmarks, coordinates, and geopolitical understanding. | 84.0 | 84.0 | n/a |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | n/a | 96.5 | 98.3 |
Remote Labor Index | 17.9 | n/a | 20.8 |
SpireClimb | 2.0 | n/a | 34.0 |
WeirdML v3 | 48.4 | n/a | 60.0 |
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 | |
| $2.00 | $12.00 | 1.0M tokens (~524 books) | $45.00 | |
| $10.00 | $50.00 | 1.1M tokens (~525 books) | $200.00 |