DeepSeek launching v4.1 flash cheaper and more capable than v4 pro
Points and comments are a snapshot, not live.
DeepSeek's V4.1 Flash surpasses V4 Pro in all key metrics, launching September 10, 2026.
DeepSeek will release V4.1 Flash around September 10, 2026 (Beijing Time). It outperforms V4 Pro on performance, cost, speed, and task completion time. After launch, Pro model requests will be routed to V4.1 Flash and billed at Flash's price. Pricing for the Flash series changes at 12:00 Beijing Time on September 10: off-peak rates are $0.003 for input cache hits, $0.15 for input cache misses, and $0.6 for output, with peak-hour prices double.
What commenters are saying
Commenters are broadly positive about DeepSeek's release cadence and value. Many use Flash models as their default for coding and agentic tasks, citing strong cost-performance trade-offs. Some emphasize the open-weights nature, allowing local hosting. One user notes V4 Pro sometimes hallucinates on complex coding tasks, while others find Flash reliable. A few commenters compare favorably to US labs' pricing and policies, with one remarking, "Imagine OpenAI/Google/Anthropic doing this!" Concerns about data privacy and Chinese surveillance are raised, but countered by arguments that US companies are similar and that models can be run locally.