DeepSeek V4 Flash Vision Exp vs DeepSeek V4.1 Flash 1017

Change models
IndexAll categories, out of 100.
60.4 ±7.2
68.5 ±2.6 best
GapIndex points below the top.
−8.1
Top
CoverageIndex weight that has results.
85%
95% best
SpeedOutput tokens per second.
74/s
261/s best
EloLMArena rating and rank.
—
1463
ReleasedMonth of release.
2026-08
2026-10
Input / 1MUS dollars per 1M input tokens.
$0.44
$0.022 best
Output / 1MUS dollars per 1M output tokens.
$1.32
$0.40 best
Batch output / 1MBatch price per 1M output tokens.
$0.33
$0.20 best
1K chats2K tokens in, 500 out, each.
$1.54
$0.244 best
1K documents20K tokens in, 1K out, each.
$10.12
$0.84 best
1K agent tasks200K tokens in, 20K out, each.
$114.40
$12.40 best
ContextMaximum tokens in one request.
1.05M
1.05M
AgenticMulti-step tasks with tools.
71.8
72.2 best
CodingCode writing and repair.
43.1
69.8 best
ReasoningLogic problems and puzzles.
68.4 best
65.0
MultimodalTasks with images and text.
—
62.2
KnowledgeFacts and expert knowledge.
65.4 best
64.2
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
70.2 best
68.3
MathMath problems.
62.2
65.4 best

Benchmarks

12 shared
Only DeepSeek V4.1 Flash 1017: 82
DeepSeek V4 Flash Vision Exp
DeepSeek V4.1 Flash 1017
79.5% best14 / 80
79.3%15 / 80
85.4%38 / 80
86.7% best31 / 80
65.1%5 / 66
77.3% best1 / 66
68.2%63 / 66
80.0% best20 / 66
71.0% best24 / 66
70.0%29 / 66
80.4%27 / 66
81.2% best26 / 66
87.8%42 / 66
93.3% best20 / 66
37.6%51 / 52
90.2% best13 / 52
70.7%38 / 52
79.2% best15 / 52
69.2%35 / 52
86.8% best4 / 52
39.0%24 / 51
40.1% best22 / 51
64.0%25 / 49
84.0% best11 / 49