GPT-5.5 is OpenAI's latest frontier model, released in April 2026 as the first fully retrained base model since GPT-4.5. It supports a 1M+ token context window with text and image inputs, delivering significantly improved token efficiency over GPT-5.4 to produce the same results at lower cost. With strengthened reasoning across coding, mathematics, research, and multi-step planning—offered at standard $5/$30 pricing—it suits a wide range of workflows.
API|VisionReasoningWeb Search|Proprietary Model
Knowledge Cutoff
2025-12-01
Input → Output Format
Context Memory
1.1MIN128KOUT
AI Performance Evaluation
Arena Overall Score
1476
±8As of 2026-05-01
Overall Rank
No.13
5,302 Votes
Arena by Ability
Hard Prompts
1491±10No.19
Expert Knowledge
1490±25No.22
Instruction Following
1465±14No.14
Conversation Memory
1487±21No.13
Creative
1430±20No.37
Coding
1509±16No.27
Math
1496±33No.7
Arena by Occupation
Creative Writing
1455±17No.17
Social Sciences
1483±19No.24
Media
1424±18No.39
Business
1482±20No.11
Healthcare
1468±32No.44
Legal
1479±31No.21
Software
1505±13No.19
Mathematics
1511±33No.6
Source:Arena Intelligence
Overall
AA Intelligence Index
60%↑21%
LiveBench
81%↑21%
Reasoning & Math
GPQA Diamond
94%↑11%
HLE
44%↑27%
LB Reasoning
88%↑19%
LB Math
96%↑22%
LB Data
81%↑28%
Coding
AA Coding Index
59%↑23%
LB Coding
82%↑10%
LB Agentic
57%↑12%
TAU2
94%↑13%
TerminalBench
61%↑27%
SciCode
56%↑14%
Language & Instructions
IFBench
76%↑13%
AA-LCR
74%↑12%
LB Language
88%↑15%
LB IF
73%↑22%
Output Speed
Standard Mode
54tok/s↓23
First Output 0.95s
Reasoning Mode
59tok/s↓28
First Output 9.17s
