team profile
Team rank#15
Score67.3
DeepSeek models · ranked
27 models · composite 0–100#ModelBest resultComposite
1BrowseComp · 83.4%672SWE-bench Verified · 76.2%623Chatbot Arena (Coding) · 1437614BrowseComp · 73.2%605Chatbot Arena (Coding) · 1491566Chatbot Arena (Coding) · 1460557MathArena Apex · 9.4%548Chatbot Arena (Coding) · 1438509Chatbot Arena (Coding) · 13375010Chatbot Arena (Coding) · 13164811Chatbot Arena (Coding) · 13064612MathArena Apex · 2.1%4613Chatbot Arena (Coding) · 12864414Arena Agent · 3.49%4315Chatbot Arena (Coding) · 14314316Chatbot Arena (Coding) · 14304317Chatbot Arena (Coding) · 13824118Chatbot Arena (Coding) · 12604119Chatbot Arena (Coding) · 12594120Chatbot Arena (Coding) · 13914021SWE-bench Pro · 15.6%3722Arena Agent · 1.19%3423MathArena Apex · 0.5%3224MathArena Apex · 0.5%3225Chatbot Arena (Coding) · 14362926Chatbot Arena (Coding) · 11062927BrowseComp · 40.1%29
DeepSeek-V4-Pro-Max
open
DeepSeek-V4
open
DeepSeek-V3.1-Thinking
open
DeepSeek-V4-Flash-Max
open
DeepSeek-V4-Pro
open
DeepSeek-V4-Flash
open
DeepSeek-v3.2-Speciale
open
DeepSeek-V3.2-Thinking
open
DeepSeek-V3
open
Deepseek-v2.5-1210
open
Deepseek-v2.5
open
DeepSeek-v3.2 (Think)
open
Deepseek-Coder-v2-0724
open
V4 Flash
open
DeepSeek-V3.2
open
DeepSeek-V3.1
open
DeepSeek-R1
open
Deepseek-v2-API-0628
open
DeepSeek-Coder-V2-Instruct
open
DeepSeek-V3-0324
open
deepseek-v3p2
open
V4 Pro
open
DeepSeek-v3.2-Exp (Think)
open
DeepSeek-v3.1 (Think)
open
DeepSeek-R1-0528
open
DeepSeek-LLM-67B-Chat
open
deepseek-v3.2-exp
open