AI BENCHY
Advertise here

#78

DeepSeek V4 Pro

DeepSeek Release: 2026-04-24 Tested on: 2026-04-29 14:46 deepseek/deepseek-v4-pro::none
(high) (none)

Score

6.2

Consistency

8.7

Reliability

7.9

Total Cost (Current Price)

$0.043

Total Output Tokens

3,903

Input Price

$0.435 / 1M

Output Price

$0.870 / 1M

Tests Correct

Wrong Tests: 11

Attempt pass rate: 48.2%

Flaky tests

3

Flaky tests had mixed outcomes across runs (at least one pass and one fail).

Response Time (avg)

14.01s

Response Time (max): 58.65s

Response Time (total): 252.12s

Wrong answer: 7 Did not follow instructions: 3 Extra formatting: 1

Run history

Tested on Score Reliability Tests Correct Total Cost Compare
2026-05-22 00:38 Suite changed 6.0 8.1 $0.046 Compare
2026-04-29 14:46 Re-test 6.2 7.9 $0.043 Current run
2026-04-24 09:19 Initial run 3.1 N/A $0.009 Compare

Run comparison

RunScoreConsistencyReliabilityTests CorrectFlaky testsTotal Output TokensTotal CostResponse Time (avg)
2026-04-29 14:46 · Re-test6.28.77.97/1833,903$0.04314.01s
2026-05-22 00:38 · Suite changed6.08.98.18/2035,347$0.04613.48s
Difference+0.1-0.2-0.2-10-1444-$0.004+529ms

These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.

Charts

Choose the first model, then click a second model to open a side-by-side page.

Total Output Tokens

Score vs Total Output Tokens

Quick Compare

DeepSeek V4 PrononevsGemini 3.1 Flash LiteminimalDeepSeek V4 PrononevsMiMo-V2-OmnimediumDeepSeek V4 PrononevsKimi K2.5mediumDeepSeek V4 PrononevsGrok 4.20mediumDeepSeek V4 PrononevsGemini 3 Flash PreviewmediumDeepSeek V4 PrononevsGemini 3.5 FlashhighDeepSeek V4 PrononevsRing-2.6-1TmediumDeepSeek V4 PrononevsGemini 3.5 Flashlow

Category Breakdown

Category Score Consistency Tests Correct
Anti-AI Tricks 3.5 8.0
Coding 7.1 3.7
Combined 9.5 10.0
Data parsing and extraction 10.0 10.0
Domain specific 5.3 10.0
General Intelligence 4.3 9.9
Instructions following 6.3 10.0
Puzzle Solving 6.0 7.1
Tool Calling 10.0 10.0

Compared models