SShogo Bench

Claude Haiku 5.5 vs GPT-5.5

Claude Haiku 5.5 leads GPT-5.5 by 0.8 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Haiku 5.5 wins 4 and GPT-5.5 wins 14 of 18 shared benchmarks. Claude Haiku 5.5 is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarClaude Haiku 5.5GPT-5.5
Agentic coding6763
Long-horizon autonomy6267
Knowledge-work agents4646
Computer use4743
Security & context learning5555

Shared benchmarks

BenchmarkClaude Haiku 5.5GPT-5.5
APEX-Agentsindex56.1%55.1%
Vending-Bench 2index$3,602$7,524
ARC-AGI89.0%95.0%
ARC-AGI-261.3%85.0%
Chess Puzzles55.0%63.0%
CritPt22.6%27.1%
DTBench95.0%96.0%
FrontierCode46.4%43.0%
FrontierMath Tier 446.3%72.5%
FrontierMath Tiers 1-375.1%85.3%
GPQA Diamond89.6%94.0%
LMCA51.1%54.3%
MirrorCode78.9%23.3%
OTIS Mock AIME98.9%100.0%
ProofBench89.0%50.0%
SciCode55.0%56.1%
SimpleBench59.2%69.0%
SimpleQA Verified23.8%63.0%

Specs and price

Claude Haiku 5.5GPT-5.5
Agent Index57.5 (#18)56.7 (#19)
Agent $/M$0.050$2.55
Input / output $/M$0.10 / $0.50$5.00 / $30.00
Context1M1.1M
ReleasedOct 7, 2026Apr 23, 2026
WeightsClosedClosed
API hosts54

Don’t pick one. Route.

Shogo can run Claude Haiku 5.5 and GPT-5.5 side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo