Navigate
AI BENCHY
Advertise here

DeepSeek V4 Flash 0423 vs GPT-5.6 Luna

The average score is effectively tied at 5.4 vs 5.4. GPT-5.6 Luna has the lower benchmark cost at $0.029 vs $0.040. GPT-5.6 Luna is faster at 1.50s vs 36.17s, with pass rates of 27.3% vs 36.4%.

Last updated at: 2026-08-20

Rank
#218
Total Output Tokens
100,726
Response Time (avg)
36.17s
Total Cost
$0.040
Rank
#221
Total Output Tokens
6,709
Response Time (avg)
1.50s
Total Cost
$0.029
Recommended model GPT-5.6 Luna

It has the best score here (5.4), while responding about 24.1x faster than DeepSeek V4 Flash 0423.

Detailed comparison

Metric DeepSeek V4 Flash 0423 DeepSeek V4 Flash 0423 none Release: 2026-04-24 GPT-5.6 Luna GPT-5.6 Luna none Release: 2026-07-09
Score 5.4 5.4
Rank #218 #221
Reliability 10.0 10.0
Consistency 8.5 8.8
Attempts 66/66 66/66
Tests Correct
Attempt pass rate 27.3% 36.4%
Flaky tests 4 3
Total Runs 66 66
Cost per result 1.146 2.357
Total Cost $0.040 $0.029
Input Price $0.089 / 1M $0.200 / 1M
Output Price $0.178 / 1M $1.200 / 1M
Total Input Tokens 240,230 101,332
Output Tokens 100,726 6,709
Reasoning Tokens 0 0
Response Time (avg) 36.17s 1.50s
Response Time (max) 247.27s 10.57s
Response Time (total) 795.83s 33.05s
Parameters 284B total (13B active) ~400B total (~17B active)
Availability Open source Closed

Model generation showcase

Hamster playing table tennis

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

#218 DeepSeek V4 Flash 0423

none
Cost
$0.004
Time
157.6s
Tokens
11,297 tok

#221 GPT-5.6 Luna

none
Cost
$0.016
Time
15.8s
Tokens
2,685 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 0423 4.2 7.4 11.1% 1 17.13s 7,279 9,717 0
GPT-5.6 Luna 3.8 7.2 22.2% 1 980ms 7,302 459 0

Quick Compare

Switch Comparison Pair