Which is better, Claude Fable 5 or GPT-6 Astra?

Refresh Models —Deals —Regulatory —Sources —

Claude Fable 5 vs GPT-6 Astra

23 SHARED BENCHMARKS

Across 23 shared benchmarks, Claude Fable 5 scores higher on 9 and GPT-6 Astra on 14. The widest gap is Terminal-Bench-Science 0.1, where GPT-6 Astra scores 64.6 against 24.7.

ANTHROPICVSOPENAI23 SHARED9–14 HEAD-TO-HEAD

BenchmarkClaude Fable 5MARGINGPT-6 Astra

AA Agentic Index56.651.5

AA Intelligence62.161.2

AA-LCR76.776.3

AA-Omniscience43.343.7

arc_agi_198.586

ARC-AGI-289.295

Artificial Analysis Coding Index76.577.1

browsecomp8891.5

critpt28.631.7

DeepSWE 1.17074.1

gdpval61.156.5

GPQA Diamond92.696.3

healthbench_professional6663.4

HLE64.554.7

MMMU-Pro84.286.9

OmniScience Accuracy65.362.6

OmniScience Non-Hallucination36.455.2

scicode60.254.1

simpleqa_verified70.775.6

TauBench V3 - Banking38.143.1

Terminal-Bench 2.18889.9

Terminal-Bench 4.044.557.7

Terminal-Bench-Science 0.124.764.6

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.