Expired75% confidenceFactTime unknown

DeepSeek V3.2 Special has achieved world-first rankings on multiple top competition benchmarks including AMC/AIME, Harvard-MIT Mathematics Tournament, and the International Mathematical Olympiad, surpassing GPT-5 and Gemini 3.0 Pro.

1

Sources

75%

Confidence

Medium-term (~90 days)

Relevance

6/1/2026

First Seen

Valid until: 8/30/2026(expired)

Sources

DeepSeek V3.2 Deep Dive: Three Major Breakthroughs in DSA Sparse Attention, Reinforcement Learning, and Agent Capabilities

bilibili网络安全基础官方教程

Related Entities

DeepSeek V3.2 Specialmodel GPT-5model Gemini 3.0 Promodel

Related Claims

UnverifiedDeepSeek V3.2 Specialty (the extended thinking version) won a gold medal at the IMO 2025 International Mathematical Olympiad.76% similar UnverifiedDeepSeek Math 7B 轻量版本在奥数和高等数学领域可追平乃至超越 GPT-4、Gemini Ultra 等闭源旗舰模型74% similar UnverifiedDeepSeek在MATH、AIME等数学竞赛基准测试上表现突出70% similar UnverifiedDeepSeek 在2024年凭借 DeepSeek-V2 和 DeepSeek-Coder 系列在国际 benchmark 上取得与 GPT-4 相当的成绩67% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/19949

API

curl https://kongchang.com/api/v1/knowledge/claims/19949

MCP

get_claim(id=19949)