AI BENCHY Compare

Google: Gemma 4 31B vs Kwaipilot: KAT-Coder-Air V2.5

Summary

Gemma 4 31B leads on average score with 6.1 vs 6.0. Gemma 4 31B has the lower benchmark cost at $0.002 vs $0.030. Gemma 4 31B is faster at 4.05s vs 6.47s, with pass rates of 47.6% vs 46.0%.

Recommended modelGemma 4 31BIt has the best score here (6.1), while costing about 17.1x less than KAT-Coder-Air V2.5 (low).

Last updated at: 2026-07-14

Metric	Gemma 4 31B Gemma 4 31B none Release: 2026-04-02 Free Available	KAT-Coder-Air V2.5 KAT-Coder-Air V2.5 low Release: 2026-07-14

Metric	Gemma 4 31B Gemma 4 31B none Release: 2026-04-02 Free Available	KAT-Coder-Air V2.5 KAT-Coder-Air V2.5 low Release: 2026-07-14
Score	6.1	6.0
Rank	#120	#124
Reliability	10.0	9.7
Consistency	10.0	7.6
Tests Correct
Attempt pass rate	47.6%	46.0%
Flaky tests	0	6
Total Runs	63	63
Cost per result	0.034	0.427
Total Cost	$0.002	$0.030
Input Price	$0.060 / 1M	$0.150 / 1M
Output Price	$0.350 / 1M	$0.600 / 1M
Total Input Tokens	20,911	42,464
Output Tokens	1,407	4,227
Reasoning Tokens	0	34,864
Response Time (avg)	4.05s	6.47s
Response Time (max)	26.13s	25.49s
Response Time (total)	76.87s	135.80s

Generation showcase

Hamster playing table tennis

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

#120 Gemma 4 31B

none

Cost: $0.001
Time: 12.8s
Tokens: 795 tok

#124 KAT-Coder-Air V2.5

low

Cost: $0.004
Time: 29.9s
Tokens: 6,008 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

Category:

Anti-AI Tricks	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	6.5	10.0	50.0%	0		1.85s	852	45	0
KAT-Coder-Air V2.5	7.3	5.8	83.3%	2		3.50s	672	422	1,548

Coding	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	5.5	10.0	33.3%	0		11.19s	8,381	735	0
KAT-Coder-Air V2.5	3.5	7.3	11.1%	1		11.06s	7,893	2,001	13,623

Combined	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	3.0	10.0	0.0%	0		0ms	0	0	0
KAT-Coder-Air V2.5	10.0	10.0	100.0%	0		25.49s	15,132	368	5,707

Data parsing and extraction	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	10.0	10.0	100.0%	0		2.25s	8,352	285	0
KAT-Coder-Air V2.5	6.5	10.0	50.0%	0		2.82s	7,776	261	1,069

Domain specific	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	7.7	10.0	66.7%	0		3.22s	903	27	0
KAT-Coder-Air V2.5	2.9	7.2	11.1%	1		4.99s	553	310	3,332

General Intelligence	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	10.0	10.0	100.0%	0		2.09s	576	117	0
KAT-Coder-Air V2.5	5.0	10.0	0.0%	0		10.11s	516	142	596

Instructions following	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	6.5	10.0	50.0%	0		2.84s	795	78	0
KAT-Coder-Air V2.5	9.8	10.0	100.0%	0		1.64s	699	89	640

Puzzle Solving	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	6.5	10.0	33.3%	0		4.23s	828	108	0
KAT-Coder-Air V2.5	3.1	4.6	22.2%	2		1.57s	696	294	796

Tool Calling	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	3.0	10.0	0.0%	0		0ms	0	0	0
KAT-Coder-Air V2.5	10.0	10.0	100.0%	0		4.47s	8,391	330	1,065

Trivia	Score	Consistency	Attempt pass rate	Flaky tests	Tests Correct	Response Time (avg)	Input Tokens	Output Tokens	Reasoning Tokens
Gemma 4 31B	3.0	10.0	0.0%	0		1.25s	224	12	0
KAT-Coder-Air V2.5	3.0	10.0	0.0%	0		19.98s	136	10	6,488

Quick Compare

Switch Comparison Pair