Navigate
AI BENCHY
Advertise here

DeepSeek: DeepSeek V4 Pro vs Google: Gemini 3.5 Flash-Lite

DeepSeek V4 Pro leads on average score with 6.9 vs 6.7. DeepSeek V4 Pro has the lower benchmark cost at $0.096 vs $0.145. Gemini 3.5 Flash-Lite (low) is faster at 2.25s vs 11.55s, with pass rates of 51.5% vs 66.7%.

Recommended modelGemini 3.5 Flash-Lite (low)Its score stays close to the best score here (6.7 vs 6.9), while responding about 5.1x faster than DeepSeek V4 Pro.

Last updated at: 2026-07-21

Metric DeepSeek V4 Pro DeepSeek V4 Pro none Release: 2026-04-24 Gemini 3.5 Flash-Lite Gemini 3.5 Flash-Lite low Release: 2026-07-21
Score 6.9 6.7
Rank #86 #95
Reliability 10.0 10.0
Consistency 8.6 8.2
Tests Correct
Attempt pass rate 51.5% 66.7%
Flaky tests 4 5
Total Runs 66 66
Cost per result 0.954 1.201
Total Cost $0.096 $0.145
Input Price $0.435 / 1M $0.300 / 1M
Output Price $0.870 / 1M $2.500 / 1M
Total Input Tokens 148,069 144,622
Output Tokens 35,551 15,302
Reasoning Tokens 0 24,971
Response Time (avg) 11.55s 2.25s
Response Time (max) 119.44s 13.50s
Response Time (total) 254.11s 49.58s

Hamster playing table tennis

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

#86 DeepSeek V4 Pro

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

#95 Gemini 3.5 Flash-Lite

low
Cost
$0.008
Time
7.7s
Tokens
3,056 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 Pro 5.6 10.0 33.3% 0 13.38s 7,275 5,500 0
Gemini 3.5 Flash-Lite 4.1 7.9 11.1% 1 672ms 8,136 463 0

Quick Compare

Switch Comparison Pair