#25
DeepSeek V4 Pro
deepseek-v4-proHelpfulness
Instruction
Following
Comprehension
Empathy
Creative
Writing
Helpfulness
0.0
Instruction Following
0.0
Comprehension
0.0
Empathy
0.0
Creative Writing
0.0
Speed
Avg 62 tok/s
Release Date
April 24, 2026
Lab
DeepSeek
Type
Open Source
Context Size
1M
Max Output Tokens
384K
Cost per 1 million tokens
$1.74 / $3.48
Model Inputs*
Text
Model Outputs*
Text
Tool Calling*
Enabled
Overall Assistant Score
An average score combining the 5 main categories.
88.10 pts
Rank #25
65th Percentile
0
Novice
33
Capable
66
Proficient
100
Expert
DeepSeek V4 Pro is DeepSeek's April 2026 open flagship for serious assistant work—long documents, careful instructions, and multi-step help. Our tests show excellent comprehension and instruction following, with helpfulness that stays on-brief through long chats. Empathy is a clear weak spot. Creative writing is competent for business drafts but our tests still prefer Claude—and especially the discontinued GPT-4.5—for voice; thinking-mode efficiency favors clear, on-brief prose over literary nuance. A top open pick when correctness and long context matter more than warmth.
Overall Score vs Price
Price axis is logarithmic (USD per 1M input tokens) · Higher score and lower price is better
Ringed point is this model
Anthropic
DeepSeek
Google
Meta
Moonshot
OpenAI
SpaceX AI
Zhipu AI
Intelligence
Overall Score · Higher is better
Speed
Output tokens per second · Higher is better
Price
USD per 1M input tokens · Lower is better
DeepSeek Models
Overall Score · Same provider
DeepSeek Family
Overall Score · Same model family
Closest Rivals
Overall Score · Nearest by overall rank