DeepSeek V4 Flash 0731 vs DeepSeek V4 Pro

Change models
IndexAll categories, out of 100.
60.6 ±3.7 best
60.3 ±3.8
GapIndex points below the top.
Top
−0.3 · no clear gap
CoverageIndex weight that has results.
85%
85%
SpeedOutput tokens per second.
93/s
96/s best
EloLMArena rating and rank.
—
1451
ReleasedMonth of release.
2026-07
2026-08
Input / 1MUS dollars per 1M input tokens.
$0.044 best
$0.90
Output / 1MUS dollars per 1M output tokens.
$0.30 best
$1.50
Batch output / 1MBatch price per 1M output tokens.
$0.144
—
1K chats2K tokens in, 500 out, each.
$0.239 best
$2.55
1K documents20K tokens in, 1K out, each.
$1.19 best
$19.50
1K agent tasks200K tokens in, 20K out, each.
$14.86 best
$210.00
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
58.7
61.4 best
CodingCode writing and repair.
62.0 best
60.0
ReasoningLogic problems and puzzles.
64.3 best
58.6
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
58.1
62.4 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
58.5 best
56.0
MathMath problems.
58.7 best
56.7

Benchmarks

71 shared
Only DeepSeek V4 Flash 0731: 14Only DeepSeek V4 Pro: 12
DeepSeek V4 Flash 0731
DeepSeek V4 Pro
91.0% best37 / 220
89.6%48 / 220
38.6%53 / 216
41.0% best43 / 216
79.7%45 / 211
80.3% best40 / 211
16.6%45 / 207
18.0% best34 / 207
94.4%50 / 203
96.7% best39 / 203
90.8%30 / 193
92.8% best14 / 193
40.4%53 / 170
49.1% best26 / 170
77.7%107 / 149
80.3% best87 / 149
88.1%39 / 144
90.1% best31 / 144
TaxEval v2Finance
70.7%88 / 143
72.1% best68 / 143
87.3%16 / 142
87.5% best13 / 142
89.9% best27 / 135
89.4%30 / 135
MMLU ProKnowledge
86.2%45 / 135
87.2% best32 / 135
46.7%35 / 132
54.5% best23 / 132
CorpFin v2Finance
61.8% best51 / 131
61.4%52 / 131
74.7% best50 / 130
49.9%73 / 130
79.0%25 / 114
80.6% best16 / 114
50.3%51 / 113
51.0% best48 / 113
MedScribeHealthcare
80.4% best59 / 111
75.1%83 / 111
MedCodeHealthcare
41.4% best54 / 108
40.5%65 / 108