SShogo Bench

Claude Fable 5.2 vs GPT-6 Astra

Claude Fable 5.2 leads GPT-6 Astra by 5.5 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Fable 5.2 wins 16 and GPT-6 Astra wins 3 of 20 shared benchmarks. Claude Fable 5.2 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarClaude Fable 5.2GPT-6 Astra
Agentic coding8680
Long-horizon autonomy7880
Knowledge-work agents6756
Computer use7767
Security & context learning6563

Shared benchmarks

BenchmarkClaude Fable 5.2GPT-6 Astra
APEX-Agentsindex80.1%64.7%
Vending-Bench 2index$9,628$15,515
FrontierSWEindex69.9%65.5%
Remote Labor Indexindex41.3%20.8%
PostTrainBenchindex44.6%41.9%
ARC-AGI98.5%98.5%
ARC-AGI-293.8%95.0%
Chess Puzzles79.0%76.0%
CritPt35.7%31.7%
DTBench99.2%97.3%
FrontierMath Tier 4100.0%97.6%
FrontierMath Tiers 1-395.1%93.7%
GPQA Diamond96.0%95.8%
LMCA70.6%64.4%
OTIS Mock AIME99.7%100.0%
ProofBench100.0%99.0%
SciCode67.7%56.5%
SimpleBench94.9%86.5%
SimpleQA Verified94.0%75.6%
WeirdML v343.1%42.2%

Specs and price

Claude Fable 5.2GPT-6 Astra
Agent Index76.4 (#1)70.9 (#4)
Agent $/M$3.21$4.62
Input / output $/M$8.00 / $40.00$10.00 / $50.00
Context1M1.1M
ReleasedNov 9, 2026Sep 4, 2026
WeightsClosedClosed
API hosts46

Don’t pick one. Route.

Shogo can run Claude Fable 5.2 and GPT-6 Astra side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo