readable AI benchmarks

Google

Gemma 4 26B A4B (Reasoning)

reasoning · open weights · Apr 2, 2026

Quality Score22.7
Value Score26.4
Reliability24.6
Cache Discount26%

Model specification

Reasoning
reasoning
Input modalities
text, image, video
Output modalities
text
Context window
256k tokens
Weights
open weights

Token prices USD per 1M tokens

Input
$0.115
Output
$0.370
Cached Input
$0.085
Cached Output
not available
Cost per task
$0.039

Capability

Intelligence
26.1
Coding
39.3
Agentic
11.0
Omniscience
-50.8
Correct
19.1%
Blended price
$0.120

Answer outcomes

Correct 19.1%Incorrect 69.9%Abstained 11.0%

Artificial Analysis benchmarks

GDPval-AA v2
13%
τ³-Banking
12%
Terminal-Bench v2.1
39%
SciCode
40%
Humanity’s Last Exam
19%
GPQA Diamond
79%
CritPt
0%
AA-Omniscience
25%
AA-LCR
62%

Similar models