Comparing 2 models
Claude Fable 5.2 vs GPT-6 Astra
Benchmarks through
Verdict
Claude Fable 5.2 places higher on 5 of the 5 tasks where both are ranked. Claude Fable 5.2 costs 1.3x less per input token. GPT-6 Astra accepts 1.1x more context.
- Choose Claude Fable 5.2 for Tool & Function Calling (#1), AI Agents (#1), Autonomous Agents (#1).
Built from the published task rankings below; ranks come from each task's own weighted benchmarks, so a win on one task says nothing about another.
Measured head to head
Benchmarks both models were run on
| Evaluation | Claude Fable 5.2 | GPT-6 Astra | Source |
|---|---|---|---|
| FrontierMath Tiers 1-3 (Epoch AI) | 95.1% ± 1.3 | 93.7% ± 1.4 | source · 2026-11-06 |
| GPQA Diamond (Epoch AI) | 93.9% ± 1.6 | 95.8% ± 1.4 | source · 2026-11-06 |
| LMArena Agent | 30.4 | 14.5 | source · 2026-11-11 |
| LMArena Agent Steerability | 33 | 14.8 | source · 2026-11-11 |
| LMArena Agent Task Outcome | 22 | 14 | source · 2026-11-11 |
| LMArena Agent Tool Hallucination | 1 | 1 | source · 2026-11-11 |
| LMArena WebDev | 1838 Elo | 1764 Elo | source · 2026-11-11 |
| OTIS Mock AIME (Epoch AI) | 99.7% ± 0.3 | 100% ± 0 | source · 2026-11-06 |
| SimpleQA Verified (Epoch AI) | 94% ± 0.8 | 75.6% ± 1.4 | source · 2026-11-06 |
| Terminal-Bench (4.1) | 65.8% | 60.3% | source · 2026-11-15 |
Values are as published by each evaluator, in its own units; rows from different evaluators are not on a common scale.
task evidence
How they place when the question changes
Lower ranks are better. “Best” marks the strongest option in this shortlist for that task.