Navigate
AI BENCHY
Advertise here

DeepSeek V4 Flash 0731 (low) vs Step 3.7 Flash (medium)

DeepSeek V4 Flash 0731 (low) leads on average score with 8.2 vs 8.0. DeepSeek V4 Flash 0731 (low) has the lower benchmark cost at $0.053 vs $0.515. Step 3.7 Flash (medium) is faster at 26.37s vs 60.18s, with pass rates of 75.8% vs 72.7%.

Last updated at: 2026-08-01

Rank
#32
Total Output Tokens
314,605
Response Time (avg)
60.18s
Total Cost
$0.053
Rank
#37
Total Output Tokens
427,572
Response Time (avg)
26.37s
Total Cost
$0.515
Recommended model DeepSeek V4 Flash 0731 (low)

It has the best score here (8.2), while costing about 9.8x less than Step 3.7 Flash (medium).

Detailed comparison

Metric DeepSeek V4 Flash 0731 DeepSeek V4 Flash 0731 low Release: 2026-08-01 Step 3.7 Flash Step 3.7 Flash medium Release: 2026-05-29
Score 8.2 8.0
Rank #32 #37
Reliability 10.0 9.9
Consistency 8.3 8.9
Tests Correct
Attempt pass rate 75.8% 72.7%
Flaky tests 5 3
Total Runs 66 64
Cost per result 0.374 3.676
Total Cost $0.053 $0.515
Input Price $0.140 / 1M $0.200 / 1M
Output Price $0.280 / 1M $1.150 / 1M
Total Input Tokens 93,336 114,062
Output Tokens 40,662 427,572
Reasoning Tokens 273,943 0
Response Time (avg) 60.18s 26.37s
Response Time (max) 438.65s 152.83s
Response Time (total) 1323.92s 580.25s

Model generation showcase

Hamster playing table tennis

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

#32 DeepSeek V4 Flash 0731

low
Invalid SVG
Cost
$0.000
Time
78.1s
Tokens
8,090 tok

#37 Step 3.7 Flash

medium
Cost
$0.006
Time
46.2s
Tokens
4,466 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

Coding Score Consistency Attempt pass rate Flaky tests Tests Correct Response Time (avg) Input Tokens Output Tokens Reasoning Tokens
DeepSeek V4 Flash 0731 7.3 5.1 77.8% 2 163.31s 5,953 782 86,028
Step 3.7 Flash 8.8 7.8 88.9% 1 27.42s 7,437 44,797 0

Quick Compare

Switch Comparison Pair