SShogo Bench

Claude Fable 5 vs GPT-6 Astra

GPT-6 Astra leads Claude Fable 5 by 6.8 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Fable 5 wins 4 and GPT-6 Astra wins 19 of 25 shared benchmarks. Claude Fable 5 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarClaude Fable 5GPT-6 Astra
Agentic coding7480
Long-horizon autonomy6980
Knowledge-work agents5256
Computer use5567
Security & context learning5963

Shared benchmarks

BenchmarkClaude Fable 5GPT-6 Astra
APEX-Agentsindex63.6%64.7%
DeepSWEindex69.9%74.1%
Vending-Bench 2index$5,680$15,515
FrontierSWEindex47.0%65.5%
Remote Labor Indexindex16.1%20.8%
GSO-Benchindex78.4%79.4%
PostTrainBenchindex37.3%41.9%
ARC-AGI98.5%98.5%
ARC-AGI-289.2%95.0%
Chess Puzzles47.0%76.0%
CritPt28.6%31.7%
DTBench98.4%97.3%
FrontierCode53.5%53.3%
FrontierMath Tier 490.2%97.6%
FrontierMath Tiers 1-387.0%93.7%
GPQA Diamond93.9%95.8%
LMCA61.2%64.4%
MirrorCode70.0%60.0%
OTIS Mock AIME100.0%100.0%
ProofBench95.0%99.0%
SciCode61.0%56.5%
SimpleBench81.9%86.5%
SimpleQA Verified70.7%75.6%
WebDev Arena16281800
WeirdML v315.6%42.2%

Specs and price

Claude Fable 5GPT-6 Astra
Agent Index64.1 (#12)70.9 (#4)
Agent $/M$4.62$4.62
Input / output $/M$10.00 / $50.00$10.00 / $50.00
Context1M1.1M
ReleasedJun 9, 2026Sep 4, 2026
WeightsClosedClosed
API hosts46

Don’t pick one. Route.

Shogo can run Claude Fable 5 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