Arena AI 中文镜像
CNDeepSeek

DeepSeek V4 Flash

DeepSeek2026-07-31● 开源 · MIT🤗 deepseek-ai/DeepSeek-V4-Flash-0731439.1万4,086

DeepSeek V4 Flash 0731 是 DeepSeek 推出的稀疏混合专家(MoE)模型,总参数 2840 亿,激活参数 130 亿。这一重训练修订版适用于编程、推理与 Agent 工作流。

模态
参数量 (MoE)
284B总
激活 13B / token
上下文
1.05Mtokens
最大输出 944K tokens
排名与分数走势
Deepseek V4 FlashDeepseek V4 Flash High (Preview)
排名
#61#77#92#1086/197/279/310/12
Elo 评分
14011425144814726/197/279/310/12
能力
工具调用 (Tools)
结构化输出
并行工具调用
视觉输入
音频输入
多服务商价格对比数据来自 OpenRouter
27 个服务商1 USD ≈ 6.6985 CNY · 更新于 2026-10-17
服务商量化精度上下文输入/M输出/M缓存读取/M近1天可用率
R
Reka
—262K¥0.14¥3.54¥0.04100.0%
BaiduBaidu限时 -86%fp81.05M¥0.40¥1.20¥0.0199.7%
DeepInfraDeepInfrafp81.05M¥0.40¥1.21¥0.10100.0%
RelaceRelacefp41.05M¥0.47¥0.94¥0.09100.0%
O
OpenInference
fp41.05M¥0.47¥0.94¥0.0999.5%
StreamLakeStreamLake限时 -80%fp81.02M¥0.59¥1.77¥0.02100.0%
S
Sail Research
fp41.05M¥0.67¥2.01¥0.20100.0%
S
Sail Research
fp41.05M¥0.67¥2.01¥0.20100.0%
DigitalOceanDigitalOcean—1.05M¥0.80¥1.59¥0.16100.0%
W
Wafer
nvfp41.05M¥0.87¥1.54¥0.84100.0%
BaseTenBaseTenfp81.05M¥0.87¥1.74¥0.1992.8%
BaseTenBaseTenfp321.05M¥0.87¥1.74¥0.1993.2%
VeniceVenice—1.00M¥0.87¥1.74¥0.19100.0%
C
CoreWeave
fp8262K¥0.87¥1.88¥0.47100.0%
I
Inceptron
fp41.05M¥0.87¥3.12¥0.1899.9%
CohereCohere—1.05M¥0.94¥1.88¥0.4799.9%
TogetherTogether—1.05M¥0.94¥1.88¥0.2099.1%
ParasailParasailfp81.05M¥0.94¥1.88¥0.33100.0%
M
Mancer 2
fp81.05M¥1.34¥4.02-99.9%
S
SiliconFlow
fp81.05M¥1.47¥4.42¥0.19100.0%
G
GMICloud限时 -35%
fp81.05M¥1.92¥5.75¥0.06100.0%
AlibabaAlibaba 时段定价—1.00M¥2.36¥7.07¥0.2499.1%
AlibabaAlibaba (US East 1)—1.00M¥2.51¥7.52¥0.25100.0%
NovitaNovita限时 -7%fp81.05M¥2.74¥8.22¥0.17100.0%
P
Phala
—1.05M¥2.95¥8.84¥0.19100.0%
AtlasCloudAtlasCloudfp41.05M¥2.95¥8.84¥0.1999.9%
CloudflareCloudflare—1.05M¥2.95¥8.84¥0.0999.9%
R
Reka
¥0.14
输入/M
¥3.54
输出/M
¥0.04
缓存/M
100.0%
可用率
BaiduBaidufp8限时 -86%
¥0.40
输入/M
¥1.20
输出/M
¥0.01
缓存/M
99.7%
可用率
DeepInfraDeepInfrafp8
¥0.40
输入/M
¥1.21
输出/M
¥0.10
缓存/M
100.0%
可用率
RelaceRelacefp4
¥0.47
输入/M
¥0.94
输出/M
¥0.09
缓存/M
100.0%
可用率
O
OpenInferencefp4
¥0.47
输入/M
¥0.94
输出/M
¥0.09
缓存/M
99.5%
可用率
StreamLakeStreamLakefp8限时 -80%
¥0.59
输入/M
¥1.77
输出/M
¥0.02
缓存/M
100.0%
可用率
S
Sail Researchfp4
¥0.67
输入/M
¥2.01
输出/M
¥0.20
缓存/M
100.0%
可用率
S
Sail Researchfp4
¥0.67
输入/M
¥2.01
输出/M
¥0.20
缓存/M
100.0%
可用率
DigitalOceanDigitalOcean
¥0.80
输入/M
¥1.59
输出/M
¥0.16
缓存/M
100.0%
可用率
W
Wafernvfp4
¥0.87
输入/M
¥1.54
输出/M
¥0.84
缓存/M
100.0%
可用率
BaseTenBaseTenfp8
¥0.87
输入/M
¥1.74
输出/M
¥0.19
缓存/M
92.8%
可用率
BaseTenBaseTenfp32
¥0.87
输入/M
¥1.74
输出/M
¥0.19
缓存/M
93.2%
可用率
VeniceVenice
¥0.87
输入/M
¥1.74
输出/M
¥0.19
缓存/M
100.0%
可用率
C
CoreWeavefp8
¥0.87
输入/M
¥1.88
输出/M
¥0.47
缓存/M
100.0%
可用率
I
Inceptronfp4
¥0.87
输入/M
¥3.12
输出/M
¥0.18
缓存/M
99.9%
可用率
CohereCohere
¥0.94
输入/M
¥1.88
输出/M
¥0.47
缓存/M
99.9%
可用率
TogetherTogether
¥0.94
输入/M
¥1.88
输出/M
¥0.20
缓存/M
99.1%
可用率
ParasailParasailfp8
¥0.94
输入/M
¥1.88
输出/M
¥0.33
缓存/M
100.0%
可用率
M
Mancer 2fp8
¥1.34
输入/M
¥4.02
输出/M
-
缓存/M
99.9%
可用率
S
SiliconFlowfp8
¥1.47
输入/M
¥4.42
输出/M
¥0.19
缓存/M
100.0%
可用率
G
GMICloudfp8限时 -35%
¥1.92
输入/M
¥5.75
输出/M
¥0.06
缓存/M
100.0%
可用率
AlibabaAlibaba 时段定价
¥2.36
输入/M
¥7.07
输出/M
¥0.24
缓存/M
99.1%
可用率
AlibabaAlibaba (US East 1)
¥2.51
输入/M
¥7.52
输出/M
¥0.25
缓存/M
100.0%
可用率
NovitaNovitafp8限时 -7%
¥2.74
输入/M
¥8.22
输出/M
¥0.17
缓存/M
100.0%
可用率
P
Phala
¥2.95
输入/M
¥8.84
输出/M
¥0.19
缓存/M
100.0%
可用率
AtlasCloudAtlasCloudfp4
¥2.95
输入/M
¥8.84
输出/M
¥0.19
缓存/M
99.9%
可用率
CloudflareCloudflare
¥2.95
输入/M
¥8.84
输出/M
¥0.09
缓存/M
99.9%
可用率
📊 Artificial Analysis 综合指数
34.3
智能指数
69.1
编程指数
41
Agent 指数