Navigate
AI BENCHY
Advertise here

AI BENCHY Compare

DeepSeek: DeepSeek V4 Pro vs StepFun: Step 3.5 Flash

Summary

DeepSeek V4 Pro vs Step 3.5 Flash benchmark comparison: DeepSeek V4 Pro leads on average score with 7.2 vs 6.9. DeepSeek V4 Pro has the lower benchmark cost at $0.030 vs $0.070. DeepSeek V4 Pro is faster at 5.30s vs 72.53s, with pass rates of 52.4% vs 56.7%.

Recommended model: DeepSeek V4 Pro - It has the best score here (7.2), while costing about 2.4x less than Step 3.5 Flash.

Last updated at: 2026-06-12

Metric DeepSeek V4 Pro DeepSeek V4 Pro none Release: 2026-04-24 Step 3.5 Flash Step 3.5 Flash medium Release: 2026-02-01
Score 7.2 6.9
Rank #61 #69
Reliability 9.9 10.0
Consistency 8.8 9.4
Tests Correct
Attempt pass rate 52.4% 56.7%
Flaky tests 3 1
Total Runs 61 60
Cost per result 0.293 0.198
Total Cost $0.030 $0.070
Input Price $0.435 / 1M $0.090 / 1M
Output Price $0.870 / 1M $0.300 / 1M
Total Input Tokens 53,078 34,431
Output Tokens 7,047 91,587
Reasoning Tokens 0 195,973
Response Time (avg) 5.30s 72.53s
Response Time (max) 23.74s 453.94s
Response Time (total) 111.39s 1015.47s

Generation showcase

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#61 DeepSeek V4 Pro

none
Invalid SVG
Cost
$0.000
Time
300.0s
Tokens
0 tok

#69 Step 3.5 Flash

medium
Cost
$0.008
Time
277.1s
Tokens
23,695 tok

Top Models by Score

Score vs Total Cost

Response Time (avg)

Score vs Response Time (avg)

Total Output Tokens

Score vs Total Output Tokens

Category Breakdown

Anti-AI Tricks Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 3.2 6.1 16.7% 2 4.02s 540 1,168 0
Step 3.5 Flash 10.0 10.0 100.0% 0 40.57s 694 20,391 24,176
Coding Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 5.6 10.0 33.3% 0 5.62s 6,795 1,123 0
Step 3.5 Flash 3.5 7.8 0.0% 0 258.38s 2,211 13,207 22,429
Combined Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 9.5 10.0 100.0% 0 23.74s 27,529 2,235 0
Step 3.5 Flash 10.0 10.0 100.0% 0 29.57s 13,638 1,176 12,984
Data parsing and extraction Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 10.0 10.0 100.0% 0 4.61s 7,568 200 0
Step 3.5 Flash 10.0 10.0 100.0% 0 15.01s 7,368 600 13,886
Domain specific Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 5.3 10.0 33.3% 0 3.72s 666 24 0
Step 3.5 Flash 5.3 7.2 44.4% 1 170.45s 673 45,350 90,436
General Intelligence Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 5.0 10.0 0.0% 0 2.05s 471 126 0
Step 3.5 Flash 5.5 10.0 0.0% 0 22.39s 509 240 3,506
Instructions following Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 6.3 5.8 66.7% 1 4.12s 627 713 0
Step 3.5 Flash 8.3 10.0 50.0% 0 4.78s 705 2,364 3,521
Puzzle Solving Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 10.0 10.0 100.0% 0 3.61s 594 442 0
Step 3.5 Flash 5.3 10.0 33.3% 0 7.22s 711 5,630 10,861
Tool Calling Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 10.0 10.0 100.0% 0 7.40s 8,105 328 0
Step 3.5 Flash 10.0 10.0 100.0% 0 11.91s 7,701 275 3,802
Trivia Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 3.0 10.0 0.0% 0 5.76s 183 688 0
Step 3.5 Flash 3.0 10.0 0.0% 0 108.45s 221 2,354 10,372

Quick Compare

Switch Comparison Pair