GPT-5.4
OpenAI
Parameters
—
Architecture
—
Released
—
License
—
About
No detailed description available for this model.
Benchmark Scores
| Benchmark | Score | Date | Standing |
|---|---|---|---|
|
AgentWorldBench MCP
general_agent
|
100.00%
|
24.06.2026 |
#1/14
better than 100%
|
|
AgentWorldBench Search
general_agent
|
96.47%
|
24.06.2026 |
#2/14
better than 92%
|
|
AgentWorldBench Terminal
general_agent
|
68.74%
|
24.06.2026 |
#7/14
better than 54%
|
|
AgentWorldBench SWE
general_agent
|
92.91%
|
24.06.2026 |
#2/14
better than 92%
|
|
AgentWorldBench Android
general_agent
|
81.37%
|
24.06.2026 |
#5/14
better than 69%
|
|
AgentWorldBench Web
general_agent
|
62.17%
|
24.06.2026 |
#3/14
better than 85%
|
|
AgentWorldBench OS
general_agent
|
88.37%
|
24.06.2026 |
#2/14
better than 92%
|
|
AgentWorldBench Overall
general_agent
|
96.35%
|
24.06.2026 |
#2/14
better than 92%
|
|
HLE (with tools)
stem_reasoning
|
68.09%
|
23.08.2026 |
#24/49
better than 52%
|
|
BrowseComp Agent Swarm
general_agent
|
78.40
|
23.08.2026 |
#2/2
better than 0%
|
|
DeepSearch QA
general_agent
|
50.75%
|
23.08.2026 |
#6/9
better than 38%
|
|
Toolathlon Verified
general_agent
|
53.49%
|
23.08.2026 |
#20/35
better than 44%
|
|
MCPMark
general_agent
|
65.29%
|
23.08.2026 |
#2/15
better than 93%
|
|
Apex-Agents
general_agent
|
43.15%
|
23.08.2026 |
#14/24
better than 43%
|
|
OSWorld-Verified
general_agent
|
80.08%
|
23.08.2026 |
#8/25
better than 71%
|
|
Terminal-Bench 2.0
coding_agent
|
91.53%
|
23.08.2026 |
#4/23
better than 77%
|
|
SWE-bench Pro
coding_agent
|
64.33%
|
23.08.2026 |
#27/77
better than 66%
|
|
SciCode (subtask)
stem_reasoning
|
88.83%
|
23.08.2026 |
#3/15
better than 86%
|
|
Humanity's Last Exam
stem_reasoning
|
71.40%
|
23.08.2026 |
#22/71
better than 70%
|
|
AIME 26
stem_reasoning
|
100.00%
|
23.08.2026 |
#1/45
better than 98%
|
|
HMMT Feb 26
stem_reasoning
|
100.00%
|
23.08.2026 |
#1/44
better than 100%
|
|
IMOAnswerBench
stem_reasoning
|
97.73%
|
23.08.2026 |
#2/30
better than 97%
|
|
GPQA Diamond
stem_reasoning
|
97.16%
|
23.08.2026 |
#8/80
better than 90%
|
|
MMMU-Pro
vision_language
|
94.97%
|
23.08.2026 |
#3/35
better than 91%
|
|
CharXiv (RQ)
document_understanding
|
66.40%
|
23.08.2026 |
#8/34
better than 79%
|
|
MathVision
vision_language
|
88.80%
|
23.08.2026 |
#2/27
better than 96%
|
|
BabyVision
vision_language
|
45.52%
|
23.08.2026 |
#11/24
better than 57%
|
|
V-Star
vision_language
|
100.00%
|
23.08.2026 |
#1/20
better than 100%
|
|
MMLU-Pro
knowledge
|
88.71%
|
26.06.2026 |
#7/48
better than 85%
|
|
CodeForces
stem_reasoning
|
90.98%
|
26.06.2026 |
#3/18
better than 88%
|
|
Apex-Shortlist (no tools)
stem_reasoning
|
80.26%
|
26.06.2026 |
#6/11
better than 50%
|
|
GDPVal-AA v2
general_agent
|
91.43%
|
26.06.2026 |
#10/38
better than 76%
|
|
BrowseComp
general_agent
|
89.44%
|
23.08.2026 |
#10/45
better than 80%
|
|
Claw-Eval Avg
coding_agent
|
94.98%
|
23.08.2026 |
#2/40
better than 97%
|
|
MCP-Atlas
general_agent
|
71.28%
|
26.06.2026 |
#32/54
better than 42%
|
|
Claw-Eval Pass^3
coding_agent
|
66.23%
|
23.08.2026 |
#6/21
better than 75%
|
What is a good AgentWorldBench MCP score? What is a good AgentWorldBench Search score? What is a good AgentWorldBench Terminal score?
No comments yet
Be the first to share your thoughts!