DeepSeek V4 Flash 0731 vs dots3-note Preview

Change models
dots3-note PreviewDots Studio
IndexAll categories, out of 100.
61.9 ±3.8
63.4 ±3.9 best
GapIndex points below the top.
−1.5 · no clear gap
Top
CoverageIndex weight that has results.
85%
95% best
SpeedOutput tokens per second.
61/s
—
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2026-07
2026-08
Input / 1MUS dollars per 1M input tokens.
$0.043
Free best
Output / 1MUS dollars per 1M output tokens.
$0.30
Free best
Batch output / 1MBatch price per 1M output tokens.
$0.144
—
1K chats2K tokens in, 500 out, each.
$0.237
Free best
1K documents20K tokens in, 1K out, each.
$1.17
Free best
1K agent tasks200K tokens in, 20K out, each.
$14.68
Free best
ContextMaximum tokens in one request.
1.05M best
512K
AgenticMulti-step tasks with tools.
59.9
63.3 best
CodingCode writing and repair.
62.8 best
60.3
ReasoningLogic problems and puzzles.
65.5
74.5 best
MultimodalTasks with images and text.
—
59.2
KnowledgeFacts and expert knowledge.
58.6
77.2 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
59.2 best
58.2
MathMath problems.
61.5
62.3 best

Benchmarks

10 shared
Only DeepSeek V4 Flash 0731: 75Only dots3-note Preview: 17
DeepSeek V4 Flash 0731
dots3-note Preview
61.4%36 / 131
76.8% best22 / 131
79.0% best24 / 114
78.4%27 / 114
82.7% best27 / 86
75.1%44 / 86
52.6%63 / 83
61.0% best31 / 83
34.8%41 / 77
52.6% best18 / 77
BrowseCompAgentic
73.2%40 / 63
83.3% best24 / 63
45.1%31 / 36
52.6% best23 / 36
NL2RepoCoding
54.2% best11 / 36
49.8%12 / 36
70.3% best18 / 23
55.6%19 / 23
88.4%4 / 10
90.9% best1 / 10