Navigate
AI BENCHY
Advertise here

Compared models

Qwen3.6 35B A3B (medium) vs Qwen3.5-27B (medium) vs DeepSeek V4 Flash 0423 (high) benchmark comparison: DeepSeek V4 Flash 0423 (high) leads on Score with 7.6. Qwen3.6 35B A3B (medium) leads on Reliability with 10.0. DeepSeek V4 Flash 0423 (high) has the lowest Total Cost at $0.031. DeepSeek V4 Flash 0423 (high) is fastest at 51.81s.

Last updated at: 2026-08-15

Rank
#148
Total Output Tokens
743,292
Response Time (avg)
58.49s
Total Cost
$0.750
Rank
#87
Total Output Tokens
614,909
Response Time (avg)
113.28s
Total Cost
$0.982
Rank
#76
Total Output Tokens
211,310
Response Time (avg)
51.81s
Total Cost
$0.031
Recommended model DeepSeek V4 Flash 0423 (high)

It has the best score here (7.6), while costing about 28.1x less than the other models in this comparison.

Detailed comparison

Metric Qwen3.6 35B A3B Qwen3.6 35B A3B medium Release: 2026-04-20 Qwen3.5-27B Qwen3.5-27B medium Release: 2026-02-24 DeepSeek V4 Flash 0423 DeepSeek V4 Flash 0423 high Release: 2026-04-24
Score 6.6 7.5 7.6
Rank #148 #87 #76
Reliability 10.0 10.0 10.0
Consistency 9.2 7.8 8.5
Attempts 66/66 66/66 66/66
Tests Correct
Attempt pass rate 57.6% 75.8% 69.7%
Flaky tests 2 6 4
Total Runs 66 66 66
Cost per result 6.212 8.217 0.446
Total Cost $0.750 $0.982 $0.031
Input Price $0.150 / 1M $0.195 / 1M $0.066 / 1M
Output Price $1.000 / 1M $1.560 / 1M $0.132 / 1M
Total Input Tokens 85,148 111,644 108,480
Output Tokens 61,736 16,113 37,229
Reasoning Tokens 681,556 598,796 174,081
Response Time (avg) 58.49s 113.28s 51.81s
Response Time (max) 817.57s 1026.43s 218.13s
Response Time (total) 1169.83s 2492.26s 1139.75s
Parameters 35B total (3B active) 27B 284B total (13B active)
Availability Open source Open source Open source

Model generation showcase

Hamster playing table tennis

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

#148 Qwen3.6 35B A3B

medium
Invalid SVG
Cost
$0.000
Time
300.0s
Tokens
0 tok

#87 Qwen3.5-27B

medium
Cost
$0.008
Time
62.0s
Tokens
3,099 tok

#76 DeepSeek V4 Flash 0423

high
Cost
$0.003
Time
93.1s
Tokens
7,926 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
Qwen3.6 35B A3B 7.7 10.0 66.7% 0 50.55s 5,051 7,929 37,223
Qwen3.5-27B 6.2 7.1 55.6% 1 160.69s 7,895 6,381 89,388
DeepSeek V4 Flash 0423 7.8 10.0 66.7% 0 50.60s 7,279 395 34,862

Quick Compare

Switch Comparison Pair