GPT-5.5

OpenAI · Closed weights · Released Apr 23, 2026 · $5 in · $30 out per 1M tokens

60Overall · rank 10

Score by job

79 · rank 6 · 5 of 5 benchmarks

Benchmark	Result	Variant	z
DeepSWE	67.0%	gpt-5.5_xhigh	0.63
Terminal-Bench	84.7%	gpt-5.5_unknown	1.29
FrontierCode	43.0%	gpt-5.5_unknown	0.35
SWE-bench Verified	80.6%	gpt-5.5-pre-release_xhigh	1.27
LMArena Coding	1495 rating	gpt-5.5-high	0.22

38 · rank 11 · 3 of 4 benchmarks

Benchmark	Result	Variant	z
APEX-Agents	55.1%	gpt-5.5_unknown	0.26
OSWorld 2.0	13.0%	gpt-5.5_xhigh	-0.54
Remote Labor Index	6.3%	gpt-5.5_unknown	-0.35

72 · rank 10 · 3 of 4 benchmarks

Benchmark	Result	Variant	z
GPQA Diamond	94.0%	gpt-5.5-pre-release_xhigh	0.98
ARC-AGI-2	85.0%	gpt-5.5_xhigh	0.97
SimpleBench	69.0%	gpt-5.5_unknown	0.36

69 · rank 7 · 5 of 5 benchmarks

Benchmark	Result	Variant	z
FrontierMath Tiers 1–3	85.3%	gpt-5.5_xhigh	1.19
FrontierMath Tier 4	72.5%	gpt-5.5_xhigh	1.32
OTIS Mock AIME	100.0%	gpt-5.5-pre-release_xhigh	1.09
ProofBench	50.0%	gpt-5.5_xhigh	0.09
LMArena Math	1486 rating	gpt-5.5	0.23

68 · rank 12 · 2 of 2 benchmarks

Benchmark	Result	Variant	z
LMArena Creative Writing	1454 rating	gpt-5.5-high	0.08
LMArena Instruction Following	1478 rating	gpt-5.5-high	0.51

31 · rank 32 · 1 of 1 benchmarks

Benchmark	Result	Variant	z
LMArena WebDev	1510 rating	gpt-5.5-xhigh (codex-harness)	-0.47