DeepSeek V4.1 Pro has no model card, no API identifier, no price and no benchmark and has never been released; it exists only as a name in a single sentence in DeepSeek's release notes for V4.1 Flash on 10 September 2026, a notice stating requests to deepseek-v4-pro would be routed to the new Flash model "until V4.1-Pro launches," and search demand for it has run for eight weeks. Ripple 65B Preview shipped yesterday, 4 November 2026, from Ricursive Intelligence under Apache 2.0, with 65 billion dense parameters and twelve trillion training tokens. DeepSeek's live models are V4.1 Flash (1017), with 1M context, 384K max output, vision support, $0.15 input / $0.60 output and $0.003 cache hit; V4 Pro (0813), with 1M context, 384K max output, no vision, $0.66 input / $1.98 output and $0.022 cache hit; and V4 Flash (0731), with 284B total parameters and 13B active, 1M context, MIT licence and $0.14 / $0.28. Calling deepseek-v4.1-pro on third-party gateways returns a model-not-found error. Ripple's launch claim is a Terminal-Bench 2.1 score