ReportingVerified newsroom

DeepSeek says new Flash AI model beats Kimi K3 on cyber, coding benchmarks - scmp.com

DeepSeek says new Flash AI model beats Kimi K3 on cyber, coding benchmarks scmp.com

DeepSeek (Google News)·Thursday, September 10, 2026 at 9:26 AM UTC

DeepSeek

Open-weight frontier lab · DeepSeek

Fetching source· Pulling live content from news.google.com…

Company hub:DeepSeek News →

Read original →

More from DeepSeek

ReportingThursday, September 10, 2026 at 10:21 AM UTC
New DeepSeek Model Launched: Long-Text Cache Hard Disk Occupancy Plummets by 88% - 36 Kr
New DeepSeek Model Launched: Long-Text Cache Hard Disk Occupancy Plummets by 88% 36 Kr
DeepSeek (Google News)
ReportingThursday, September 10, 2026 at 9:47 AM UTC
China’s DeepSeek Launches Smaller, Faster AI Model - Caixin Global
China’s DeepSeek Launches Smaller, Faster AI Model Caixin Global
DeepSeek (Google News)
ReportingThursday, September 10, 2026 at 9:23 AM UTC
DeepSeek’s New Architecture Slashes Agentic Costs by 80% - forkast.news
DeepSeek’s New Architecture Slashes Agentic Costs by 80% forkast.news
DeepSeek (Google News)
ReportingThursday, September 10, 2026 at 9:21 AM UTC
DeepSeek V4.1 Flash Launched: Unleash Pro-Level AI Capabilities at Affordable Flash Version Pricing - 36 Kr
DeepSeek V4.1 Flash Launched: Unleash Pro-Level AI Capabilities at Affordable Flash Version Pricing 36 Kr
DeepSeek (Google News)
ReportingThursday, September 10, 2026 at 9:11 AM UTC
V4.1 Flash Outperforms Across the Board: Why Developers Still Criticize DeepSeek – Its Gap Is Not Capability, But Software Engineering Mindset - 36 Kr
V4.1 Flash Outperforms Across the Board: Why Developers Still Criticize DeepSeek – Its Gap Is Not Capability, But Software Engineering Mindset 36 Kr
DeepSeek (Google News)
ReportingThursday, September 10, 2026 at 8:56 AM UTC
DeepSeek V4.1 Flash: Stronger, Faster, More Accessible - Geopolitechs
DeepSeek V4.1 Flash: Stronger, Faster, More Accessible Geopolitechs
DeepSeek (Google News)

DeepSeek says new Flash AI model beats Kimi K3 on cyber, coding benchmarks - scmp.com

New DeepSeek Model Launched: Long-Text Cache Hard Disk Occupancy Plummets by 88% - 36 Kr

China’s DeepSeek Launches Smaller, Faster AI Model - Caixin Global

DeepSeek’s New Architecture Slashes Agentic Costs by 80% - forkast.news

DeepSeek V4.1 Flash Launched: Unleash Pro-Level AI Capabilities at Affordable Flash Version Pricing - 36 Kr

V4.1 Flash Outperforms Across the Board: Why Developers Still Criticize DeepSeek – Its Gap Is Not Capability, But Software Engineering Mindset - 36 Kr

DeepSeek V4.1 Flash: Stronger, Faster, More Accessible - Geopolitechs