GPT-5.4 vs GPT-5.4 Pro

Change models
GPT-5.4OpenAI
IndexAll categories, out of 100.
64.8 ±2.5
73.3 ±8.7 best
GapIndex points below the top.
−8.5 · no clear gap
Top
CoverageIndex weight that has results.
95% best
60%
SpeedOutput tokens per second.
38/s best
21/s
EloLMArena rating and rank.
1452
—
ReleasedMonth of release.
2026-03
2026-03
Input / 1MUS dollars per 1M input tokens.
$2.50 best
$30.00
Output / 1MUS dollars per 1M output tokens.
$15.00 best
$180.00
Batch output / 1MBatch price per 1M output tokens.
$7.50 best
$90.00
1K chats2K tokens in, 500 out, each.
$12.50 best
$150.00
1K documents20K tokens in, 1K out, each.
$65.00 best
$780.00
1K agent tasks200K tokens in, 20K out, each.
$800.00 best
$9,600.00
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
62.1
70.5 best
CodingCode writing and repair.
61.8
—
ReasoningLogic problems and puzzles.
69.0
80.4 best
MultimodalTasks with images and text.
62.1
—
KnowledgeFacts and expert knowledge.
66.4
67.2 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
63.4
—
MathMath problems.
68.0
75.3 best

Benchmarks

15 shared
Only GPT-5.4: 84Only GPT-5.4 Pro: 3
GPT-5.4
GPT-5.4 Pro
93.3%19 / 217
94.6% best8 / 217
23.4%23 / 214
30.0% best10 / 214
74.0%24 / 131
83.3% best19 / 131
93.7%26 / 126
94.5% best21 / 126
84.2%20 / 122
92.2% best12 / 122
96.4%25 / 116
98.3% best8 / 116
45.1%43 / 91
46.3% best40 / 91
Chess PuzzlesReasoning
48.0%19 / 90
59.0% best10 / 90
78.6%22 / 90
82.5% best17 / 90
52.1%20 / 77
58.7% best8 / 77
49.0%23 / 72
58.5% best20 / 72
47.6%4 / 69
50.0% best3 / 69
BrowseCompAgentic
82.7%27 / 63
89.3% best11 / 63
27.1%7 / 55
37.5% best3 / 55
39.8%29 / 53
42.7% best25 / 53