GPT-5 vs GPT-5 mini

Change models
GPT-5OpenAI
IndexAll categories, out of 100.
53.5 ±3.4 best
48.0 ±3.9
GapIndex points below the top.
Top
−5.5
CoverageIndex weight that has results.
95%
95%
SpeedOutput tokens per second.
77/s
77/s
EloLMArena rating and rank.
1406 best
1374
ReleasedMonth of release.
2025-08
2025-08
Input / 1MUS dollars per 1M input tokens.
$1.25
$0.25 best
Output / 1MUS dollars per 1M output tokens.
$10.00
$2.00 best
Batch output / 1MBatch price per 1M output tokens.
$5.00
$1.00 best
1K chats2K tokens in, 500 out, each.
$7.50
$1.50 best
1K documents20K tokens in, 1K out, each.
$35.00
$7.00 best
1K agent tasks200K tokens in, 20K out, each.
$450.00
$90.00 best
ContextMaximum tokens in one request.
400K
400K
AgenticMulti-step tasks with tools.
49.8 best
41.3
CodingCode writing and repair.
51.3 best
49.0
ReasoningLogic problems and puzzles.
54.3 best
46.8
MultimodalTasks with images and text.
55.9
57.2 best
KnowledgeFacts and expert knowledge.
57.4 best
45.9
MultilingualTasks in many languages.
—
51.8
InstructionTasks with strict prompt rules.
56.5
—
MathMath problems.
54.8 best
51.6

Benchmarks

40 shared
Only GPT-5: 20Only GPT-5 mini: 10
GPT-5
GPT-5 mini
86.2% best72 / 220
75.0%110 / 220
91.4% best59 / 203
86.7%72 / 203
86.0% best17 / 149
81.8%80 / 149
TaxEval v2Finance
73.4%49 / 143
75.2% best22 / 143
85.9%25 / 142
86.6% best21 / 142
9.9% best73 / 137
4.4%90 / 137
85.6% best45 / 135
80.3%66 / 135
MMLU ProKnowledge
86.5% best41 / 135
82.2%78 / 135
65.7% best67 / 132
54.3%77 / 132
CorpFin v2Finance
61.1% best58 / 131
60.2%70 / 131
20.1% best98 / 130
14.2%107 / 130
9.6% best63 / 127
5.8%68 / 127
65.9% best66 / 116
61.5%73 / 116
MedScribeHealthcare
83.7% best44 / 111
80.6%58 / 111
MedCodeHealthcare
49.6% best21 / 108
43.0%49 / 108
65.5%41 / 97
66.9% best31 / 97
AIMEMath
93.4% best14 / 95
91.5%24 / 95
MedQAHealthcare
96.3% best4 / 95
96.1%5 / 95
50.1% best33 / 94
21.6%79 / 94
Chess PuzzlesReasoning
36.0% best45 / 93
24.0%58 / 93