priorsBuilder
Support

Settings

Theme

Explore themes

Mode

Library

Looks borrowed from benchmark sites and products. Themes with a light and a dark version follow your mode.

Bring your own theme

Upload a Priors theme file, or have your agent make one: copy the prompt, tell it the look you want, and upload what it returns. With the Priors skill installed it already knows the format.

Custom themes stay in this browser. They travel inside shared chart links and a page's theme, so others see them as you do.

For the builder, Explore and model pages on this device. Published pages keep the theme their owner picked.

Mode

Google · released 2026-09-02

Gemini 3.8 Flash

Measured on 238 benchmarks across 3 reasoning efforts.

30.9

#25 of 67 models · best at high

Reasoning efforts in this ranking

high30.9#60 of 149
medium30.1#66 of 149
low22.6#98 of 149

Strengths

  • Gemini 3.8 Flash
  • Gemini 4 Carbon (#1)
  • Field median
bestField medianFrontierCode 2.0-…Terminal-Bench v4…AA-Briefcase v1.1AutomationBench-AATerminal-Bench-Sc…HLE-DiamondCritPt v2 Hard Su…AA-Omniscience v1…AA-Omniscience v1…GDP.pdf
  1. FrontierCode 2.0-AA15%18.7%#32 of 67
  2. Terminal-Bench v4-AA10%19.7%#35 of 67
  3. AA-Briefcase v1.115%1203#44 of 67
  4. AutomationBench-AA5%59.9%#24 of 67
  5. Terminal-Bench-Science-AA15%10.0%#23 of 67
  6. HLE-Diamond10%33.1%#17 of 67
  7. CritPt v2 Hard Subset10%46.5%#18 of 67
  8. AA-Omniscience v1.1 Accuracy10%50.5%#11 of 67
  9. AA-Omniscience v1.1 Non-hallucination5%44.9%#45 of 67
  10. GDP.pdf5%21.0%#24 of 67

Bars show where Gemini 3.8 Flash sits among the 67 ranked models on each benchmark (right is best); the tick is the field median.

Cost and speed

against 67 ranked models

Cost per task

#31 cheapest of 50

$0.31at high

Time per task

#28 quickest of 49

14 minat high

Output speed

#22 fastest of 59

115 tok/sat high

Time to first token

#39 quickest of 59

22sat high

Output tokens per task

#34 fewest of 53

99kat high

Time per task: Artificial Analysis · computed, not timed: output tokens per Intelligence Index task divided by output speed, without tool time. Method changed Nov 4, 2026.

Token price

$0.429 per 1M tokens at API prices$0.017 on Google AI Pro

48.7% cache read · 51.1% fresh input · 0.00% cache write · 0.25% output

Measured from T3 Code usage in antigravity, Oct 30 to Nov 5, medium confidence. One app’s users, not everyone’s. $0.110 at the standard 97 / 2.5 / 0.5 mix.

This ranking’s benchmarks

10 of 10 measured

agentic

AutomationBench-AA

245 results
  1. #39medium of 24560.9%
  2. #44high of 24559.9%
  3. #122low of 24536.5%

Leader: Gemini 4 Carbon · 77.8%

agentic

AA-Briefcase v1.1

250 results
  1. #104high of 2501203
  2. #110medium of 2501183
  3. #131low of 2501062

Leader: Sonnet 5.5 · 1813

reasoning

CritPt v2 Hard Subset

116 results
  1. #42high of 11646.5%
  2. #43medium of 11645.0%
  3. #61low of 11633.5%

Leader: GPT-6 Astra · 72.0%

coding

FrontierCode 2.0-AA

113 results
  1. #58high of 11318.7%
  2. #67medium of 11316.5%
  3. #84low of 1138.1%

Leader: Opus 5.5 · 46.4%

economics

GDP.pdf

243 results
  1. #57medium of 24322.8%
  2. #68high of 24321.0%
  3. #87low of 24318.6%

Leader: Muse Wonder Full · 32.6%

reasoning

HLE-Diamond

152 results
  1. #38high of 15233.1%
  2. #44medium of 15229.1%
  3. #55low of 15224.3%

Leader: Gemini 4 Carbon · 69.1%

knowledge

AA-Omniscience v1.1 Accuracy

122 results
  1. #26high of 12250.5%
  2. #29medium of 12249.6%
  3. #34low of 12248.3%

Leader: Fable 5.1 · 68.4%

knowledge

AA-Omniscience v1.1 Non-hallucination

122 results
  1. #68medium of 12248.5%
  2. #71high of 12244.9%
  3. #89low of 12231.9%

Leader: Gemini 4 Carbon · 95.8%

reasoning

Terminal-Bench-Science-AA

126 results
  1. #51medium of 12610.5%
  2. #54high of 12610.0%
  3. #61low of 1268.1%

Leader: Gemini 4 Carbon · 64.8%

coding

Terminal-Bench v4-AA

241 results
  1. #78high of 24119.7%
  2. #79medium of 24119.7%
  3. #106low of 24110.1%

Leader: Kimi K3.1 · 67.2%

All 238 benchmarks

show 228 more