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

Update notices

A note when a benchmark one of your indexes uses gets a new version, with a one-click upgrade.

Google · released 2026-11-18

Gemini 4 Neon

Measured on 46 benchmarks across 4 reasoning efforts.

Compare

61.4

#3 of 81 models · best at xhigh

Reasoning efforts in this ranking

xhigh61.4#6 of 188
high59.6#9 of 188
medium54.9#20 of 188
low44.1#47 of 188

Strengths

  • Gemini 4 Neon
  • Fable 5.2 (#1)
  • Field median
bestField medianFrontierCode 2.0-AATerminal-Benchv4-AAAutomationBench-AAAA-Briefcase v1.1GDP.pdfTerminal-Bench-Science-AACritPt v2 HardSubsetAA-Omnisciencev1.1 AccuracyHLE-DiamondAA-Omnisciencev1.1Non-hallucination
  1. FrontierCode 2.0-AA15%44.4%#4 of 81
  2. Terminal-Bench v4-AA10%63.6%#4 of 81
  3. AutomationBench-AA5%79.3%#1 of 81
  4. AA-Briefcase v1.115%1700#5 of 81
  5. GDP.pdf5%28.0%#10 of 81
  6. Terminal-Bench-Science-AA15%70.0%#3 of 81
  7. CritPt v2 Hard Subset10%60.5%#11 of 81
  8. AA-Omniscience v1.1 Accuracy10%60.5%#9 of 81
  9. HLE-Diamond10%71.3%#3 of 81
  10. AA-Omniscience v1.1 Non-hallucination5%85.6%#5 of 81

Bars show where Gemini 4 Neon sits among the 81 ranked models on each benchmark (right is best); the tick is the field median.

By category

  • Gemini 4 Neon
  • Fable 5.2 (#1)
  • Field median
p25p50p75topField medianCodingAgents & computeruseProfessional workScience & researchReasoning & puzzlesKnowledge &accuracyVisionWriting & design
  1. Coding7 benchmarks90%#2 of 81
  2. Agents & computer use20 benchmarks93%#1 of 81
  3. Professional work10 benchmarks95%#1 of 81
  4. Science & research6 benchmarks90%#2 of 80
  5. Reasoning & puzzles3 benchmarks86%#5 of 79
  6. Knowledge & accuracy2 benchmarks98%#1 of 80
  7. Vision3 benchmarks83%#7 of 67
  8. Writing & design2 benchmarks44%#28 of 53

Across 53 results on every current benchmark that measured Gemini 4 Neon, not only this ranking’s. The percentage is its average standing among the 81 models in this ranking that took the same benchmarks: 100% beat them all, 0% trailed them all. A benchmark in two categories counts in both; a category needs two benchmarks.

Cost and speed

against 81 ranked models

Output speed

#28 fastest of 66

110 tok/sat xhigh

Time to first token

#52 quickest of 66

57sat xhigh

Output tokens per task

#61 fewest of 73

142kat xhigh

API default effort: high, per Artificial Analysis.

Token price

$0.637 per 1M tokens at API prices$0.026 on Google AI Ultra 5x (assumed same as Flash)

85.0% cache read · 14.5% fresh input · 0.00% cache write · 0.50% output

Assumed mix: no measured T3 Code usage for this model.

Not measured for Gemini 4 Neon: cost per task, time per task (no published API price yet).

Availability

Subscriptions

API value per $
  • Google AI Pro$19.99/monot measured
  • Google AI Ultra 5xtentative$99.99/mo25x
  • Google AI Ultra 20xtentative$199.99/mo49x

API providers

2
ProviderIn / out per 1MTok/sFirst
  • Google$3.00 / $15.0010130.4s
  • Google AI Studio$3.00 / $15.00

OpenRouter, Artificial Analysis

This ranking’s benchmarks

10 of 10 measured

Cost per task is what a subscriber pays: the model’s plan price, a share of API price. Models no plan sells are at API price.

Agents & computer use

AutomationBench-AA

282 results
  1. #1xhigh of 28279.3%
  2. #2high of 28279.2%
  3. #4medium of 28277.8%
  4. #12low of 28276.0%

Leader: this model · 79.3%

Professional work

AA-Briefcase v1.1

287 results
  1. #8xhigh of 2871700
  2. #20high of 2871632
  3. #55medium of 2871518
  4. #108low of 2871297

Leader: Sonnet 5.5 · 1815

Science & research

CritPt v2 Hard Subset

159 results
  1. #30low of 15962.0%
  2. #33medium of 15961.5%
  3. #38xhigh of 15960.5%
  4. #44high of 15959.0%

Leader: Fable 5.2 · 76.5%

Coding

FrontierCode 2.0-AA

156 results
  1. #9xhigh of 15644.4%
  2. #12high of 15643.3%
  3. #30medium of 15636.9%
  4. #60low of 15626.9%

Leader: Fable 5.2 · 49.1%

Professional work

GDP.pdf

281 results
  1. #1high of 28133.4%
  2. #2medium of 28132.8%
  3. #9low of 28131.4%
  4. #35xhigh of 28128.0%

Leader: this model · 33.4%

Knowledge & accuracy

HLE-Diamond

196 results
  1. #5xhigh of 19671.3%
  2. #7high of 19668.4%
  3. #16medium of 19663.6%
  4. #32low of 19654.2%

Leader: Fable 5.2 · 80.2%

Knowledge & accuracy

AA-Omniscience v1.1 Accuracy

169 results
  1. #29xhigh of 16960.5%
  2. #33high of 16959.4%
  3. #39medium of 16957.9%
  4. #40low of 16956.6%

Leader: Fable 5.2 · 79.1%

Knowledge & accuracy

AA-Omniscience v1.1 Non-hallucination

169 results
  1. #9high of 16986.1%
  2. #10xhigh of 16985.6%
  3. #12medium of 16983.8%
  4. #18low of 16981.2%

Leader: Gemini 4 Carbon · 95.8%

Science & research

Terminal-Bench-Science-AA

167 results
  1. #7xhigh of 16770.0%
  2. #12high of 16765.7%
  3. #25medium of 16754.8%
  4. #53low of 16727.1%

Leader: GPT-6.2 Astra · 78.6%

Agents & computer use

Terminal-Bench v4-AA

278 results
  1. #5xhigh of 27863.6%
  2. #9high of 27861.6%
  3. #32medium of 27854.5%
  4. #74low of 27832.8%

Leader: Kimi K3.1 · 67.2%

All 46 benchmarks

show 36 more