DeepSeek V4.1 Flash (DeepSeek) AI Model: Benchmark Scores, Pricing & VRAM Specs | ModelAnalysis.si

DeepSeek • OPEN WEIGHTSDeepSeek V4.1 FlashDeepSeek October 2026 ultra-fast MoE reasoning model breaking hyperscaler pricing with $0.15/1M input rates and 180 tok/s inference speeds.

            Metric Dimension
            Verified Value
            Evaluation Benchmark
          
            Chatbot Arena Elo
            1365 pts
            LMSYS Live Battles
          
            Coding Index
            94.5 / 100
            SWE-bench Verified
          
            Agentic Tool-Use
            93 / 100
            Multi-Turn Tool Calls
          
            Input Rate
            $0.15 / 1M
            Commercial Token API
          
            Output Rate
            $0.60 / 1M
            Generation Billing
          
            Context Horizon
            256k
            Maximum Token Window
          
      Dedicated Cloud DeploymentDeploy DeepSeek V4.1 Flash on high-performance dedicated NVIDIA GPUs at 30–60% off hyperscaler rates.
Deploy on RunPod Spot Cloud →