GPT-5 vs GPT-5.4 mini

Change models
GPT-5OpenAI
IndexAll categories, out of 100.
53.5 ±3.4
53.9 ±2.4 best
GapIndex points below the top.
−0.4 · no clear gap
Top
CoverageIndex weight that has results.
95%
100% best
SpeedOutput tokens per second.
77/s
111/s best
EloLMArena rating and rank.
1406
1412 best
ReleasedMonth of release.
2025-08
2026-03
Input / 1MUS dollars per 1M input tokens.
$1.25
$0.75 best
Output / 1MUS dollars per 1M output tokens.
$10.00
$4.50 best
Batch output / 1MBatch price per 1M output tokens.
$5.00
$2.25 best
1K chats2K tokens in, 500 out, each.
$7.50
$3.75 best
1K documents20K tokens in, 1K out, each.
$35.00
$19.50 best
1K agent tasks200K tokens in, 20K out, each.
$450.00
$240.00 best
ContextMaximum tokens in one request.
400K
400K
AgenticMulti-step tasks with tools.
49.8
55.3 best
CodingCode writing and repair.
51.3
54.6 best
ReasoningLogic problems and puzzles.
54.3 best
51.7
MultimodalTasks with images and text.
55.9 best
54.3
KnowledgeFacts and expert knowledge.
57.4 best
53.3
MultilingualTasks in many languages.
—
57.5
InstructionTasks with strict prompt rules.
56.5 best
52.1
MathMath problems.
54.8 best
53.3

Benchmarks

38 shared
Only GPT-5: 22Only GPT-5.4 mini: 38
GPT-5
GPT-5.4 mini
86.2%72 / 220
86.9% best69 / 220
28.5% best84 / 216
28.1%88 / 216
78.2% best68 / 211
77.0%80 / 211
5.7%77 / 207
10.0% best64 / 207
91.4% best59 / 203
88.9%66 / 203
85.4%73 / 193
87.5% best53 / 193
40.3% best54 / 170
37.5%66 / 170
TaxEval v2Finance
73.4% best49 / 143
71.2%80 / 143
84.8%55 / 143
93.4% best29 / 143
85.9% best25 / 142
81.5%60 / 142
73.1%31 / 140
73.3% best28 / 140
9.9%73 / 137
18.9% best66 / 137
85.6% best45 / 135
83.1%57 / 135
MMLU ProKnowledge
86.5% best41 / 135
84.6%60 / 135
65.7% best67 / 132
63.7%69 / 132
20.9%81 / 132
25.6% best71 / 132
CorpFin v2Finance
61.1% best58 / 131
60.9%62 / 131
20.1%98 / 130
48.0% best77 / 130
9.6%63 / 127
17.8% best59 / 127
65.9%66 / 116
75.1% best59 / 116