download dots

Google · tested in Taskade Genesis

Gemini 3.6 Flash in Taskade

Gemini 3.6 Flash from Google is not in the Taskade model menu right now. Its test results stay on this page. We last tested it on Aug 3, 2026 in TSK-1, our app-building test.

Status
Paused
Your own key
Google Vertex AI
Last test
Context window
1.05Mtokens
Max output
66Ktokens
Released
Knowledge cutoff
Where it runs
3services

TSK-1 scorecard for the Gemini family

TSK Score

Gemini family, across real apps made in Taskade Genesis

TSK Score 44 out of 100
Gemini family on the four TSK-1 axesInterfaceTaskMemoryAdapt
Interfaceready to share
Limited
Taskcompletes your task
Emerging
Memorykeeps your data
Emerging
Adapthandles follow-up edits
Emerging

An early Gemini result never opened: nothing to see and nothing to use.

More

The next one built and published. That is real progress, confirmed by opening the app rather than taking the model's word for it.

Gemini on TSK-1

Published scores

Benchmark scores that Google and test sites published for Gemini 3.6 Flash. Taskade did not re-run them. Lab-reported marks a score from Google itself.

Published benchmark scores for Gemini 3.6 Flash, with sources
BenchmarkScore
SWE-Bench Proresolve rate · Antigravity · Google DeepMindLab-reported58.7%
Terminal-Benchaccuracy · v2.1 · Terminus 2 · Google DeepMindLab-reported78%
OSWorld-Verifiedsuccess rate · Google DeepMindLab-reported83%
More published scores (7)
Published benchmark scores for Gemini 3.6 Flash, with sources: More published scores (7)
BenchmarkScore
DeepSWEresolve rate · v1.1 · high reasoning · Google DeepMindLab-reported49%
MLE-Benchaverage position score · Partial 30 · Google DeepMindLab-reported63.9
GDPval-AAElo · v2 · Google DeepMindLab-reported1,421
CharXiv Reasoningaccuracy · no tools · Google DeepMindLab-reported85.2%
CharXiv Reasoningaccuracy · with tools · Google DeepMindLab-reported89.4%
GDM-MRCRaccuracy · v2 · 128k average, 8-needle · Google DeepMindLab-reported91.8%
GDM-MRCRaccuracy · v2 · 1M pointwise, 8-needle · Google DeepMindLab-reported54%
How Gemini builds apps in TSK-1, our own test

Scores collected by models.dev (MIT). Each source links the page that published the score.

Versions tested

Open a version to read its findings.

Gemini 3.6 FlashTested Aug 20262 findings

An early Gemini result never opened: nothing to see and nothing to use. The next one built and published. That is real progress, confirmed by opening the app rather than taking the model's word for it.

Aug 1, 2026

The finished build never opened: three missing pieces, no light-and-dark styling, and links pointing at a machine nobody else can reach. No app shipped.

Aug 3, 2026

Built and published this time, a clear step up. It still used too many resources to qualify as the best value in the test.

See every Gemini result

Model facts

The token limits are the most the model accepts. Taskade can run it with lower limits.

Last updated
Input types
Text, Images, Video, Audio, PDF
Output types
Text
Reasoning
Yes
Tool calling
Yes
File attachments
Yes
Structured output
Yes
Open weights
No
Reasoning effort
MinimalLowMediumHigh

Published specs from the model maker, via models.dev (MIT).

Run Gemini 3.6 Flash with your own key

Gemini 3.6 Flash is not in the Taskade model menu right now. You can still pick it with your own key in a custom agent.

Key providers for Gemini 3.6 Flash, and the plan each one needs
Key providerPlan
Google Vertex AIPro and up

Where it runs

Services outside Taskade that serve Gemini 3.6 Flash, from models.dev. Each link opens in a new tab. In Taskade, your own key works only through the key providers on this page.

What to pick instead

  1. Leave TSK-Auto on, and Taskade picks a current model for each request.
  2. To pick a model yourself, choose one from the model menu.
  3. To compare, open the models page and read the test results of the current models.

Gemini in the docs

Where Gemini ranks

TSK Score out of 100.

  1. 1GPT88 out of 100
  2. 2Claude75 out of 100
  3. 2DeepSeek75 out of 100
  4. 2Kimi75 out of 100
  5. 2Grok75 out of 100
  6. 6GLM63 out of 100
  7. 7Qwen50 out of 100
  8. 8Gemini44 out of 100

Compare Gemini

Gemini vs ClaudeGemini vs DeepSeekGemini vs GPTKimi vs Gemini

FAQ

Can I still pick Gemini 3.6 Flash in Taskade?

Only with your own key. Gemini 3.6 Flash is not in the Taskade model menu right now, and TSK-Auto picks a current model for each request. In a custom agent, you can still pick Gemini 3.6 Flash with your own provider key.

Has Taskade tested Gemini 3.6 Flash?

Yes. We last tested Gemini 3.6 Flash on Aug 3, 2026 in the TSK-1 Benchmark. An early Gemini result never opened: nothing to see and nothing to use. The next one built and published. That is real progress, confirmed by opening the app rather than taking the model's word for it.

Can I use my own API key for Gemini 3.6 Flash?

Yes. In a custom agent, you can pick Gemini 3.6 Flash with your own key from Google Vertex AI (Pro and up). Your provider bills the model use to your own account.

Who makes Gemini 3.6 Flash?

Google makes Gemini 3.6 Flash. Taskade runs frontier models from top AI labs.