DeepSeek V4.1-Flash Matches Top Closed Models on Coding at a Fraction of the Cost
DeepSeek released V4.1-Flash on 10 September 2026 with MIT-licensed weights, positioning it as a cost-efficient open-source model for coding agents. On the DeepSWE v1.1 benchmark, it scores 74.2, nearly matching Claude Opus 5 (74.0) and GPT-5.6 Sol (73.0), according to DeepSeek's own release figures. Its off-peak cached-input rate of $0.003 per million tokens makes it dramatically cheaper than rivals, with one worked example showing $0.15 versus $25 for Claude Opus 5 on a 100-request agent loop. However, the model significantly trails on hard reasoning tasks, scoring 36.8 on Humanity's Last Exam compared to Claude Opus 5's 56.3, making it less suitable for complex single-answer problems. From 14 September 2026, all V4-Pro API requests are automatically rerouted to V4.1-Flash and billed at the lower Flash rate.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in