Gemini 3.6 Flash
Google · Released 21 Jul 2026 · gemini-3.6-flash
$0.75
Input price / 1M tokens
#18 of 61, cheapest first
$3.75
Output price / 1M tokens
#21 of 61, cheapest first
1.05M
Context window
#11 of 71
185
Output tokens / second
#12 of 61
58.7%
SWE-bench Pro
#9 of 16
49%
DeepSWE v1.1
#17 of 17
78%
Terminal-Bench 2.1
#12 of 17
83%
OSWorld-Verified
#3 of 8
Summary
Gemini 3.6 Flash is a proprietary model from Google, released on 21 Jul 2026.
At $0.75 input and $3.75 output per million tokens, it is #18 of 61 on input price, cheapest first.
Its 1.05M-token context window ranks #11 of 71.
Measured output speed is 185 tokens per second, #12 of 61.
Previous-generation Flash model with better token efficiency than 3.5 Flash, for code generation, agentic execution and everyday multimodal tasks.
Benchmarks · 8 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| SWE-bench Prointernal Antigravity harness, official methodology | 58.7% | #9 of 16 | Google Gemini 3.6 Flash evaluation reportSelf-reported | |
| DeepSWE v1.1public leaderboard, high reasoning | 49% | #17 of 17 | Datacurve DeepSWE leaderboard, cited in Google's evaluation report | |
| Terminal-Bench 2.1Terminus 2 harness | 78% | #12 of 17 | Google Gemini 3.6 Flash evaluation reportSelf-reported | |
| MLE-BenchPartial-30 subset, average position score, k=2 | 63.9% | – | Google Gemini 3.6 Flash evaluation reportSelf-reported | |
| GDPVal-AA v2Artificial Analysis leaderboard, Jul 2026 | 1,421 | Elo-style | – | Artificial Analysis, cited in Google's evaluation report |
| OSWorld-Verifiedaverage of 5 runs, max 100 steps | 83% | #3 of 8 | Google Gemini 3.6 Flash evaluation reportSelf-reported | |
| CharXiv Reasoningno tools (89.4% with search and code execution) | 85.2% | #2 of 6 | Google Gemini 3.6 Flash evaluation reportSelf-reported | |
| MRCR v2 (8-needle)128k average (54.0% at 1M pointwise) | 91.8% | #2 of 13 | Google Gemini 3.6 Flash evaluation reportSelf-reported |
Self-reported means the lab ran the test itself. A rank appears only where several labs report the same version of a benchmark.
Similar models
Models with the most benchmarks in common with Gemini 3.6 Flash, and the closest scores.
SWE-bench Pro
#9 of 16Harder multi-file engineering tasks
- Gemini 3.6 Flash58.7
- Gemini 3.5 Flash53.9
- Gemini 3.5 Flash-Lite54.2
- Qwen3.8-27B61.7
- GPT-5.6 Luna62.7
DeepSWE v1.1
#17 of 17Long coding tasks in real repos
Terminal-Bench 2.1
#12 of 17Terminal tasks, earlier set
OSWorld-Verified
#3 of 8Using a desktop computer
Compare Gemini 3.6 Flash side by side
Pre-filled with the two closest models. Swap any of them.
Sources
- price, cache, batch, free tier: ai.google.dev/gemini-api/docs/pricing, checked 2 Oct 2026
- context, max output, modalities, api id: ai.google.dev/gemini-api/docs/models/gemini-3.6-flash, checked 2 Oct 2026
- release date, GA status: ai.google.dev/gemini-api/docs/changelog, checked 2 Oct 2026
- knowledge cutoff: deepmind.google/models/model-cards/gemini-3-6-flash/, checked 2 Oct 2026
- benchmarks, launch price: storage.googleapis.com/deepmind-media/gemini/gemini_3-6_flash_model_evaluation.p, checked 2 Oct 2026
- introductory price now applies to 3.6 Flash: storage.googleapis.com/deepmind-media/gemini/gemini_3-7_flash_model_evaluation.p, checked 2 Oct 2026
- launch post: blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash, checked 2 Oct 2026
- output speed: artificialanalysis.ai/models/gemini-3-6-flash, checked 2 Oct 2026
Building on Gemini 3.6 Flash?
Get the architecture right before the bill arrives.
We size caching, routing and fallbacks for your workload on a free call, and tell you where a cheaper model is good enough.