DeepSeek V4 Pro arrives US-hosted at $0.75 per research task
- Source
- x.com
- Date
DeepSeek V4 Pro, hosted in the U.S., is now available in Perplexity Computer. We evaluated it against other models on WANDR. It scored 0.359 at $0.75 per task, 62% cheaper than the next model on the cost-performance frontier.
Context
Perplexity's places DeepSeek V4 Pro at a low-cost point on its WANDR research comparison. The chart shows a score of 0.359 at $0.75 per task; the next configuration on the cost-performance curve, GPT-5.6 Terra, scores 0.399 at $1.98. The advertised 62% saving therefore comes with a lower score. It describes a cheaper performance tier, not an equal-score replacement for that next model. The company also announced that its Computer deployment of V4 Pro is US-hosted. The post and chart establish both claims.
- eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
- benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
Checking sign-in…
Loading comments…