DeepSeek V4 Pro 0813 vs DeepSeek V4 Flash 0731

Change models
IndexAll categories, out of 100.
64.0 ±4.6 best
60.6 ±3.7
GapIndex points below the top.
Top
−3.4 · no clear gap
CoverageIndex weight that has results.
85%
85%
SpeedOutput tokens per second.
93/s
93/s
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-08
2026-07
Input / 1MUS dollars per 1M input tokens.
$1.32
$0.044 best
Output / 1MUS dollars per 1M output tokens.
$3.96
$0.30 best
Batch output / 1MBatch price per 1M output tokens.
$1.98
$0.144 best
1K chats2K tokens in, 500 out, each.
$4.62
$0.239 best
1K documents20K tokens in, 1K out, each.
$30.36
$1.19 best
1K agent tasks200K tokens in, 20K out, each.
$343.20
$14.86 best
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
58.5
58.7 best
CodingCode writing and repair.
65.3 best
62.0
ReasoningLogic problems and puzzles.
67.5 best
64.3
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
62.0 best
58.1
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
62.7 best
58.5
MathMath problems.
64.3 best
58.7

Benchmarks

41 shared
Only DeepSeek V4 Pro 0813: 5Only DeepSeek V4 Flash 0731: 44
DeepSeek V4 Pro 0813
DeepSeek V4 Flash 0731
91.7% best34 / 220
91.0%37 / 220
98.6% best30 / 203
94.4%50 / 203
82.4% best72 / 149
77.7%107 / 149
TaxEval v2Finance
73.1% best52 / 143
70.7%88 / 143
87.5% best13 / 142
87.3%16 / 142
61.3%42 / 137
61.4% best40 / 137
92.4% best16 / 135
89.9%27 / 135
MMLU ProKnowledge
87.0% best35 / 135
86.2%45 / 135
90.0% best39 / 132
89.0%42 / 132
CorpFin v2Finance
65.4% best27 / 131
61.8%51 / 131
82.3% best30 / 130
74.7%50 / 130
59.4%45 / 127
63.9% best36 / 127
93.3%43 / 116
94.8% best38 / 116
MedScribeHealthcare
80.2%61 / 111
80.4% best59 / 111
MedCodeHealthcare
42.5% best51 / 108
41.4%54 / 108
40.9% best38 / 98
30.3%61 / 98
52.9% best29 / 94
33.6%66 / 94
Chess PuzzlesReasoning
46.0% best29 / 93
32.0%52 / 93
64.6% best42 / 93
57.5%49 / 93
26.8%43 / 89
28.1% best41 / 89