SShogo Bench

Claude Sonnet 5 vs GPT-5.6 Sol

GPT-5.6 Sol leads Claude Sonnet 5 by 9.4 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Sonnet 5 wins 3 and GPT-5.6 Sol wins 18 of 21 shared benchmarks. Claude Sonnet 5 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarClaude Sonnet 5GPT-5.6 Sol
Agentic coding5771
Long-horizon autonomy6572
Knowledge-work agents4346
Computer use4053
Security & context learning5158

Shared benchmarks

BenchmarkClaude Sonnet 5GPT-5.6 Sol
APEX-Agentsindex54.5%51.4%
DeepSWEindex53.8%72.7%
Vending-Bench 2index$6,378$9,619
CursorBenchindex34.1%41.7%
GSO-Benchindex37.3%76.5%
Chess Puzzles45.0%68.0%
CritPt16.9%32.3%
DTBench92.5%96.0%
FrontierCode42.7%47.5%
FrontierMath Tier 429.3%82.9%
FrontierMath Tiers 1-365.6%89.1%
GeoBench74.0%73.0%
GPQA Diamond90.5%93.5%
LMCA50.0%59.2%
MirrorCode63.3%20.0%
OTIS Mock AIME94.7%100.0%
ProofBench77.0%83.0%
SciCode54.3%57.8%
SimpleBench60.6%71.7%
SimpleQA Verified33.7%69.7%
WebDev Arena15371617

Specs and price

Claude Sonnet 5GPT-5.6 Sol
Agent Index52.4 (#32)61.8 (#14)
Agent $/M$0.92$1.85
Input / output $/M$2.00 / $10.00$4.00 / $20.00
Context1M1.1M
ReleasedJun 30, 2026Aug 7, 2026
WeightsClosedClosed
API hosts55

Don’t pick one. Route.

Shogo can run Claude Sonnet 5 and GPT-5.6 Sol side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo