readable AI benchmarks (simplified)

readable AI benchmarks

Google

Gemini 3.5 Flash-Lite

reasoning · closed weights · Jul 21, 2026

Gemini 3.5 Flash-Lite has a Quality Score of 43.4, ranking 158th among 571 scored models. Per 1M tokens, pricing is $0.300 input, $2.50 output, and $0.030 cached input. It accepts text, image, video, audio, and pdf input, outputs text, and has a 1M token context window. Its Reliability Score is 52.6, ranking 101st among 573 scored models, above average (total average: 36.5). Its Value Score is 34.6, around average compared with the total Value average of 31.4.

Quality Score43.4coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Score34.6
Factual reliability52.6
Cache Discount90%

Model specification

Reasoning
reasoning
Input modalities
text, image, video, audio, pdf
Output modalities
text
Context window
1M tokens
Weights
closed weights

Token prices USD per 1M tokens

Input
$0.300
Output
$2.50
Cached Input
$0.030
Cache write
not available
Cost per task
$0.193

Capability

Intelligence
22.2
Coding
not available
Agentic
not available
Omniscience
5.2
Correct
29.5%
Blended price
$0.331

Answer outcomes

Fully graded outcomes (Correct + Incorrect): 53.7%

Correct 29.5%Incorrect 24.2%Partial / not attempted 46.3%

Artificial Analysis benchmarks

GDPval-AA v2
24%
τ³-Banking
18%
SciCode
41%
Humanity’s Last Exam
19%
GPQA Diamond
84%
CritPt
0%
AA-Omniscience
53%
AA-LCR
76%

Similar models

Support me! Patreon