AI Benchmark Model
DeepSeek V4 Flash 0420 (Non-reasoning)
Benchmark scores, pricing, speed, and model comparisons for DeepSeek V4 Flash 0420 (Non-reasoning).
Key scores
DeepSeek V4 Flash is a non-reasoning AI model released by DeepSeek on April 24, 2026. The model holds an intelligence index of 18.9. Performance metrics across various benchmarks include a GPQA score of 0.716, an HLE score of 0.078, and an IFBench score of approximately 0.472. Additionally, the model achieved a score of 0.417 on LCR, 0.341 on TerminalBench Hard, and 0.944 on TAU2. Pricing for the model is structured at $0.09 per 1 million input tokens and $0.20 per 1 million output tokens, resulting in a blended rate of $0.12 per 1 million tokens. Specific data regarding coding and math indices, output speed, and time to first token remain currently unknown.
Benchmark results
6 tracked benchmark rows for DeepSeek V4 Flash 0420 (Non-reasoning).
TAU2 Other
94.4% GPQA Reasoning
71.6% IFBench Other
47.2% LCR Other
41.7% TerminalBench Hard Other
34.1% HLE Other
7.8%