True
DeepSeek has released "V4.1 Flash," a low-cost addition to its AI model lineup that the company advertises as cutting output costs by 70 percent. The headline figure, however, is measured against the flagship V4-Pro โ not against the outgoing V4-Flash the new model replaces, a comparison against which the price cut narrows to roughly 9 percent, according to a cross-check of the company's listed API pricing.
The release itself, the listed prices and the size of the reductions all match primary disclosures exactly, according to a verification review covering seven primary sources. As stated in DeepSeek's own materials, the 70 percent discount measures V4.1 Flash's output cost against V4-Pro, the company's flagship LLM. The figure stands on those terms, but it does not describe the savings available to users already on the budget tier.
Measured against V4-Flash, the previous-generation model it succeeds, V4.1 Flash's output-price reduction comes to approximately 9 percent. The contrast frames the marketing pitch: substantial savings for customers trading down from the flagship, far slimmer ones for those upgrading from the budget model the new release replaces. The distinction matters for cost-sensitive deployments, where the choice between the two comparisons can make the same product look like either a steep discount or a marginal refresh.
Reviewers found no primary sources substantiating separate claims about CAPTCHA or authentication blocks surrounding the release. Output pricing was corroborated through a cross-check of the company's listed rates, which matched the figures cited in the announcement.
On the strength of seven primary sources, the claim that DeepSeek has released V4.1 Flash with output costs 70 percent below those of its flagship V4-Pro is True.