DeepSeek V4 Flash — benchmark results

DeepSeek's efficient open V4 Flash MoE model (13B active). Provider: DeepSeek. Released 2026-04-23. Access: Open.

Unified ELO 1741 ± 12, rank #211 of 1776 rated models, from 160 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
ATRBench23.7TSAcc Default (%) (self-reported)100
IMO-AnswerBench91.1Score (self-reported)96.2
Vellum - LiveCodeBench91.6Pass@1 (%)95.5
Beyond the All-in-One Agent57.33avg. (self-reported)90.9
OpenCompass Agent - Tool Use46.7Score (%)90.9
OpenCompass LLM - Agent46.7Score (%)90.9
PlanBench-XL63.08Accuracy (%) (self-reported)88.9
PinchBench91.46Success Rate (%)86
ClawProBench61.47Final Score (self-reported)85.7
Frontier Security Agents - Cybench Hard86.4Success rate (%)85.7
FullHome63Goal SR (VirtualHome) (self-reported)85.7
MLX Benchmark V2 - MLX Core86.17Accuracy (%)85.7

Interactive version: theaggregate.ai/model?slug=deepseek-v4-flash · How the rankings work · Data refreshed daily, snapshot 2026-07-22.