#31
DeepSeek V4 Flash
DeepSeek · Open Weight · Reasoning · 1M context
56/ 100 composite
Released 2026-07-31
Benchmark Scores by Category
Coding
79.0
Knowledge
86.2
Reasoning
78.7
Agentic
56.9
Math
94.8
Scores based on weighted benchmarks: SWE-bench, LiveCodeBench (Coding); HLE, MMLU-Pro (Knowledge); ARC-AGI-2, LongBench v2 (Reasoning); Terminal-Bench, BrowseComp (Agentic); FrontierMath, AIME (Math).
API Pricing
Input
$0.14
per 1M tokens
Output
$0.28
per 1M tokens
Value Score
200.0
score per $1 output
Estimate Your Costs
Use our cost calculators to estimate monthly spend based on your specific workload: