Navigate
AI BENCHY
Advertise here

AI BENCHY Compare

DeepSeek: DeepSeek V4 Pro vs StepFun: Step 3.7 Flash

Summary

DeepSeek V4 Pro vs Step 3.7 Flash benchmark comparison: Step 3.7 Flash leads on average score with 7.7 vs 7.6. DeepSeek V4 Pro has the lower benchmark cost at $0.157 vs $0.341. Step 3.7 Flash is faster at 15.74s vs 77.20s, with pass rates of 66.7% vs 68.3%.

Recommended model: Step 3.7 Flash - It has the best score here (7.7), while responding about 4.9x faster than DeepSeek V4 Pro.

Last updated at: 2026-06-17

Metric DeepSeek V4 Pro DeepSeek V4 Pro high Release: 2026-04-24 Step 3.7 Flash Step 3.7 Flash low Release: 2026-05-29
Score 7.6 7.7
Rank #41 #39
Reliability 9.3 10.0
Consistency 7.0 8.4
Tests Correct
Attempt pass rate 66.7% 68.3%
Flaky tests 8 4
Total Runs 63 63
Cost per result 1.742 2.840
Total Cost $0.157 $0.341
Input Price $0.435 / 1M $0.200 / 1M
Output Price $0.870 / 1M $1.150 / 1M
Total Input Tokens 38,726 40,101
Output Tokens 6,334 289,325
Reasoning Tokens 159,151 0
Response Time (avg) 77.20s 15.74s
Response Time (max) 416.76s 124.75s
Response Time (total) 1621.17s 330.63s

Generation showcase

Hamster playing table tennis

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

#41 DeepSeek V4 Pro

high
Cost
$0.023
Time
257.6s
Tokens
14,870 tok

#39 Step 3.7 Flash

low
Invalid SVG
Cost
$0.004
Time
25.3s
Tokens
3,072 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 5.7 5.9 58.3% 2 25.70s 536 149 3,214
Step 3.7 Flash 8.7 7.9 91.7% 1 4.02s 756 10,896 0
Coding Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 6.1 4.6 66.7% 2 243.00s 5,090 383 84,580
Step 3.7 Flash 8.2 7.2 88.9% 1 9.46s 7,437 18,685 0
Combined 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 38.17s 14,060 454 5,836
Step 3.7 Flash 10.0 10.0 100.0% 0 7.98s 13,683 6,426 0
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 25.03s 7,690 274 2,166
Step 3.7 Flash 7.3 5.8 83.3% 1 2.29s 7,398 2,667 0
Domain specific Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 3.6 7.2 22.2% 1 151.46s 569 4,404 50,391
Step 3.7 Flash 5.3 7.2 44.4% 1 43.31s 828 104,487 0
General Intelligence 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 8.83s 471 115 1,013
Step 3.7 Flash 3.4 9.3 0.0% 0 7.00s 525 4,604 0
Instructions following Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 7.8 6.6 83.3% 1 8.73s 627 66 2,726
Step 3.7 Flash 9.8 10.0 100.0% 0 1.58s 735 1,857 0
Puzzle Solving Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 6.9 4.9 77.8% 2 56.85s 591 178 2,563
Step 3.7 Flash 5.5 9.9 33.3% 0 1.84s 756 3,564 0
Tool Calling Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Pro 9.8 10.0 100.0% 0 15.92s 8,909 295 701
Step 3.7 Flash 10.0 10.0 100.0% 0 3.25s 7,746 1,360 0
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 34.01s 183 16 5,961
Step 3.7 Flash 3.0 10.0 0.0% 0 124.75s 237 134,779 0

Quick Compare

Switch Comparison Pair