Gemini 3.5 Flash vs Gemini 3.6 Flash

  • Gemini 3.5 Flash and Gemini 3.6 Flash are level on LMArena’s text leaderboard (1482 and 1480).
  • Gemini 3.6 Flash is cheaper: $0.75 / $3.75 per million input / output tokens, against $1.50 / $9 for Gemini 3.5 Flash.
  • On the 10 Vals AI benchmarks both have taken, Gemini 3.5 Flash wins 5 and Gemini 3.6 Flash wins 5 (Gemini 3.5 Flash: Tax Agent Bench, Legal Research Bench, MedCode).
LMArena text rating over time
14601470148014901500May 26Jun 26Aug 26Sep 26Oct 26
LatestGemini 3.5 Flash 1482Gemini 3.6 Flash 1480

Side by side

Models (2/6)
  • Gemini 3.5 Flash
  • Gemini 3.6 Flash
BenchmarkGemini 3.5 FlashGemini 3.6 Flash
The model
LabGoogleGoogle
Released2026-05-192026-07-21
WeightsProprietaryProprietary
Context window1M1M
Input, per 1M tokens$1.50$0.75
Output, per 1M tokens$9$3.75
LMArena, people’s votes
LMArena TextMillions of people compare two anonymous models side by side and vote for the better answer1482#18 of 3511480#20 of 351
LMArena WebDevPeople ask two anonymous models to build the same web app, try both and vote for the better one1499#63 of 1241538#49 of 124
LMArena VisionSide-by-side votes on prompts that include an image1308#13 of 1461299#20 of 146
LMArena CodingLMArena’s text votes on prompts that ask for code1493#33 of 3461491#34 of 346
LMArena MathLMArena’s text votes on math prompts1504#12 of 3391507#11 of 339
LMArena Creative writingLMArena’s text votes on stories, poems and other creative writing1470#17 of 3491466#21 of 349
LMArena Hard promptsLMArena’s text votes on the hardest prompts1488#23 of 3511486#25 of 351
LMArena Instruction followingLMArena’s text votes on prompts with specific instructions to follow1467#33 of 3511467#31 of 351
Vals AI, coding
Terminal-Bench 4.0Frontier-difficulty tasks an agent completes in a real terminal6.06%#45 of 55—
SWE-benchReal GitHub issues from open-source projects78.80%#27 of 8579.60%#25 of 85
Vibe Code BenchBuilding complete web apps from a plain-language brief, graded on whether the app works48.68%#60 of 11564.00%#50 of 115
LiveCodeBenchFresh competitive programming problems, collected after models’ training cut-offs to avoid contamination87.60%#12 of 14388.08%#8 of 143
Vals AI, finance
Tax Agent BenchAgentic research into US corporate tax questions21.65%#47 of 7618.04%#53 of 76
Vals AI, law
Legal Research BenchAgentic US legal research30.77%#49 of 8525.00%#60 of 85
Vals AI, medicine
MedCodeAssigning the right medical billing codes from clinical notes55.83%#6 of 10753.15%#11 of 107
Vals AI, reasoning
GPQA DiamondGraduate-level questions in biology, physics and chemistry that experts answer well and search does not92.68%#14 of 13893.43%#7 of 138
MMLU-ProHarder, reasoning-focused multiple choice across 14 subjects, from law to engineering89.52%#10 of 13889.28%#13 of 138
MMMUCollege-level questions that need both an image and text88.27%#8 of 9388.38%#7 of 93
Terminal-Bench ScienceExpert-written scientific research workflows, run by an agent in a terminal5.71%#21 of 424.29%#28 of 42

Add models in the compare tool.

Questions

Which is better, Gemini 3.5 Flash or Gemini 3.6 Flash?

Gemini 3.5 Flash and Gemini 3.6 Flash are level on LMArena’s text leaderboard (1482 and 1480). Gemini 3.6 Flash is cheaper: $0.75 / $3.75 per million input / output tokens, against $1.50 / $9 for Gemini 3.5 Flash. On the 10 Vals AI benchmarks both have taken, Gemini 3.5 Flash wins 5 and Gemini 3.6 Flash wins 5 (Gemini 3.5 Flash: Tax Agent Bench, Legal Research Bench, MedCode).

Which is cheaper, Gemini 3.5 Flash or Gemini 3.6 Flash?

Gemini 3.6 Flash is cheaper: $0.75 / $3.75 per million input / output tokens, against $1.50 / $9 for Gemini 3.5 Flash.

Sources

Every number on this page comes from these sources; models without a published number show none. Download the data. Cite as “Nexus Explorer, nexus.inc/models”.