GPT-5.5

OpenAI

Parameters

—

Architecture

—

Released

—

License

—

GPT

About

No detailed description available for this model.

Benchmark Scores

Benchmark Score Date
Humanity's Last Exam
stem_reasoning
74.43%
17.06.2026
HLE (with tools)
stem_reasoning
73.94%
17.06.2026
CritPt (no tools)
stem_reasoning
83.60%
17.06.2026
AIME 26
stem_reasoning
98.85%
17.06.2026
HMMT Feb 26
stem_reasoning
98.70%
17.06.2026
GPQA Diamond
stem_reasoning
98.68%
17.06.2026
SWE-bench Pro
coding_agent
73.25%
17.06.2026
NL2Repo
coding_agent
62.15%
17.06.2026
DeepSWE 1.1
coding_agent
93.66%
17.06.2026
ProgramBench
coding_agent
98.40%
17.06.2026
Terminal-Bench 2.1 (Terminus-2)
coding_agent
92.92%
17.06.2026
Terminal-Bench 2.1 (Best Reported Harness)
coding_agent
100.00%
17.06.2026
PostTrainBench
coding_agent
52.22%
17.06.2026
SWE-Marathon
coding_agent
22.45%
17.06.2026
MCP-Atlas
general_agent
85.77%
17.06.2026
Tool Decathlon
general_agent
91.02%
17.06.2026
Kimi Code Bench V2
coding_agent
100.00%
23.08.2026
Program Bench
coding_agent
100.00%
23.08.2026
MLS-Bench-Lite
coding_agent
37.93%
23.08.2026
Kimi Claw 24/7 Bench
agentic
100.00%
23.08.2026
MCPMark-Verified
agentic
100.00%
23.08.2026
FrontierSWE
coding_agent
73.65%
17.06.2026
HMMT Nov 25
stem_reasoning
100.00%
17.06.2026

Related Models