Storia in 5 fonti

Read DeepSeek's claim, then read DeepSeek's table

DeepSeek's experimental multimodal model wins three of eleven benchmarks against Anthropic's Opus-4.8, on DeepSeek's own published table.

Raccontata da

Confronto fonti

5 prospettive sulla stessa storia

AI · summaries

thenextweb.comStai leggendo1 g fa

Read DeepSeek's claim, then read DeepSeek's table

DeepSeek released V4-Flash-Vision-Exp claiming agent parity with Opus-4.8; it wins 3 of 11 benchmarks, loses 12 on NL2Repo. At $0.87 per million tokens versus Anthropic's $50, price is decisive, but performance gaps on repository-scale work remain critical for enterprise agents.

originale

Timeline cronologica

venerdì 21 agosto 2026·cryptobriefing.com
DeepSeek unveils AI model challenging Anthropic's dominance
DeepSeek launches a new AI model challenging Anthropic's lead. Anthropic's best AI model by September 2026 at 90.5% YES.
venerdì 21 agosto 2026·cryptobriefing.com
DeepSeek unveils AI model that approaches Anthropic's performance at a fraction of the cost
DeepSeek's V4-Flash model matches Anthropic's Claude Opus 4.8 on reasoning and coding benchmarks while costing roughly 99% less per output.

Read DeepSeek's claim, then read DeepSeek's table

Confronto fonti

Read DeepSeek's claim, then read DeepSeek's table

Timeline cronologica

DeepSeek unveils AI model challenging Anthropic's dominance

DeepSeek unveils AI model that approaches Anthropic's performance at a fraction of the cost

Deepseek releases experimental Flash vision model that rivals Opus 4.8 on agent benchmarks

DeepSeek debuts multimodal language model competitive with Opus 4.8 - SiliconANGLE

DeepSeek Enters the Multimodal AI Race with Experimental Vision Model

DeepSeek unveils AI model challenging Anthropic's dominance

Read DeepSeek's claim, then read DeepSeek's table

DeepSeek Enters the Multimodal AI Race with Experimental Vision Model

Deepseek releases experimental Flash vision model that rivals Opus 4.8 on agent benchmarks

DeepSeek debuts multimodal language model competitive with Opus 4.8 - SiliconANGLE