modelgrep
D

DeepSeek: DeepSeek V3

deepseek/deepseek-chat

ToolsJSON
Use via OpenRouter ↗
Intelligence
Design Elo
1168
3D
Speed
26
260th fastest
Latency
595ms
first token
Input price
$0.200
102nd cheapest
Context
131K
16K max out

How it compares

Faster than16%
of all ranked models
Cheaper than66%
of all ranked models

Overview

DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions. Pre-trained on nearly 15 trillion tokens, the reported evaluations...

Benchmarks

independent · via OpenRouter
Design Arena · Elo9,090 tournaments
3D
1168
Website
1165
codecategories
1160
UI Component
1151
Data Viz
1142
Game Dev
1123
svg
1033

Providers & pricing (3)

ProviderIn $/MOut $/MUptime
StreamLake$0.200$0.80099.8%
DeepInfrafp4$0.320$0.89094.8%
Novitafp8$0.400$1.30100%

Specifications

Context window131K
Max output16K
Knowledge cutoffJul 2024
Input modalitiestext
Output modalitiestext
Prompt caching
Cache read price
ModeratedNo

DeepSeek V3 FAQ

How much does DeepSeek V3 cost?

DeepSeek V3 costs $0.200 per million input tokens and $0.800 per million output tokens via OpenRouter, making it 102nd cheapest of 298 paid models.

How fast is DeepSeek V3?

DeepSeek V3 generates around 26 tokens per second with 595ms time-to-first-token (p50), the 260th fastest tracked model.

What is DeepSeek V3's context window?

DeepSeek V3 supports a 131K-token context window and can output up to 16K tokens. It accepts text input.

Compare head-to-head