SmophyAI

Published daily · Next update 02:00 UTC

Real data comparison

DeepSeek: DeepSeek V4 Flash 0423 vs StepFun: Step 5 Preview

On quality, StepFun: Step 5 Preview leads (43.7 vs 34.3). On price, DeepSeek: DeepSeek V4 Flash 0423 is cheaper ($0.66 vs $1.42 per 1M). In real production usage, DeepSeek: DeepSeek V4 Flash 0423 is used more (rank #4 vs #6 of 388). Across 3 real task categories where both models have classified traffic, DeepSeek: DeepSeek V4 Flash 0423 handles more of the traffic in 3 of them.

  • DeepSeek: DeepSeek V4 Flash 0423
  • StepFun: Step 5 Preview
Quality (Intelligence Index)34.3 / 43.7
Price ($/1M blended)$0.66 / $1.42
Average usage share12.02% / 8.16%
Best provider uptime100.0% / 100.0%
Context window1,048,576 / 1,048,576
Quality per dollar52.0 pts/$ / 30.7 pts/$

Head-to-head by real task

DeepSeek: DeepSeek V4 Flash 0423 leads 3 of 3 shared task categories by real classified traffic share; StepFun: Step 5 Preview leads 0. Not a benchmark score - this is what developers actually route to each model for, per OpenRouter's task classification.

TaskDeepSeek: DeepSeek V4 Flash 0423StepFun: Step 5 Preview
DevOps5.6%3.1%
Repo Scanning4.5%3.1%
Summarization6.0%4.9%

Usage share over time

24.64%18.48%12.32%6.16%0.00%Share of usageJul 5Aug 27Oct 20
  • DeepSeek: DeepSeek V4 Flash 0423
  • StepFun: Step 5 Preview

Frequently asked questions

DeepSeek: DeepSeek V4 Flash 0423 vs StepFun: Step 5 Preview: which is better quality?

StepFun: Step 5 Preview scores higher on Artificial Analysis's Intelligence Index (43.7 vs 34.3).

DeepSeek: DeepSeek V4 Flash 0423 vs StepFun: Step 5 Preview: which is cheaper?

DeepSeek: DeepSeek V4 Flash 0423 is cheaper at $0.66/1M vs $1.42/1M.

DeepSeek: DeepSeek V4 Flash 0423 vs StepFun: Step 5 Preview: which is used more?

DeepSeek: DeepSeek V4 Flash 0423 has higher real OpenRouter usage share (rank #4 vs #6 of 388).

DeepSeek: DeepSeek V4 Flash 0423 vs StepFun: Step 5 Preview: which wins more real-world task categories?

DeepSeek: DeepSeek V4 Flash 0423 handles more traffic in 3 of 3 shared task categories where both models have classified OpenRouter usage.

Source: OpenRouter (openrouter.ai/rankings), as of October 21, 2026.

Benchmark scores: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).

Methodology: www.smophy.ai/benchmark/methodology

Token counts originate from each provider's own tokenizer and are not directly comparable across providers.