GPT-5.4 Mini brings the core capabilities of GPT-5.4 to a faster, more efficient form factor optimized for high-throughput workloads. It runs over 2× faster than GPT-5 Mini while approaching GPT-5.4's performance on coding and reasoning benchmarks, and supports text and image inputs with full tool use, web search, and function calling. With a 400K-token context window, it delivers reliable instruction following and multi-step reasoning at significantly reduced cost, making it well-suited for chat applications, coding assistants, and agent workflows operating at scale.
API|VisionReasoningWeb SearchFile|Proprietary Model
Knowledge Cutoff
2025-08-31
Input → Output Format
Context Memory
400KIN128KOUT
AI Performance Evaluation
Arena Overall Score
1457
±6As of 2026-04-23
Overall Rank
No.27
11,237 Votes
Arena by Ability
Hard Prompts
1479±8No.28
Expert Knowledge
1487±19No.20
Instruction Following
1444±10No.32
Conversation Memory
1473±13No.23
Creative
1417±15No.41
Coding
1505±12No.26
Math
1433±22No.47
Arena by Occupation
Creative Writing
1435±12No.32
Social Sciences
1465±14No.37
Media
1426±13No.33
Business
1468±13No.18
Healthcare
1454±22No.58
Legal
1462±21No.32
Software
1493±9No.26
Mathematics
1460±24No.28
Source:Arena Intelligence
Overall
AA Intelligence Index
38%↓1%
LiveBench
34%↓26%
ForecastBench
55%↓4%
Reasoning & Math
GPQA Diamond
82%↑1%
HLE
17%↑0%
LB Reasoning
22%↓38%
LB Math
37%↓37%
LB Data
47%↓2%
Coding
AA Coding Index
38%↑3%
LB Coding
75%↑1%
LB Agentic
17%↓26%
TAU2
37%↓37%
TerminalBench
34%↑3%
SciCode
44%↑3%
Language & Instructions
IFBench
65%↑8%
AA-LCR
61%↑0%
Hallucination (HHEM)
5.5%↓5%
Factual (HHEM)
95%↑5%
LB Language
42%↓30%
LB IF
19%↓27%
Output Speed
Standard Mode
172tok/s↑90
First Output 0.48s
Reasoning Mode
180tok/s↑92
First Output 8.59s
