DeepSeek v4 0813 Pro is DeepSeek's latest and best model. It's beaten models from Google and Meta in benchmarks.

DeepSeek V4 Flash delivers competitive AI performance for budget-conscious users. Running a full test suite costs just $72, well below Gemini models.

DeepSeek released V4 Flash on July 31, scoring 79.0% on SWE-bench Verified — within 1.8 points of Claude Opus 4.7 — at roughly $0.09-0.14 per million input tokens against Claude Opus 4.8's $5. It ...

Myrto Uzuni scored the second hat trick in club history to lead Austin FC to a 3-0 win over Club Puebla in Leagues Cup Sunday ...

Gotham has lost only one game since the NWSL’s monthlong pause for the men’s World Cup in June, falling 2-0 to San Diego on ...

Higher benchmark scores don't mean lower cost. Qwen 3.8-Max and Claude Opus 5 both show it — and cost per successful task is ...

Lionel Messi scored his first two goals since the World Cup and took the career scoring lead in the Leagues Cup as Inter ...

The Phoenix Suns and veteran forward Dillon Brooks agreed to a $73 million, three-year extension on Wednesday that will keep ...

Microsoft shifts most security scanning to an in‑house AI model, cutting costs and routing only the hardest work to frontier ...

DeepSeek V4-Flash 0731 beats V4-Pro-Preview on all nine published agent benchmarks after retraining alone. Full benchmark ...

DeepSeek's newly launched V4-Flash model is over 100 times cheaper to run than Anthropic's Claude Fable 5, offering unmatched cost-efficiency among globally recognized AI models.