Gemini 3.5 Flash-Lite
Google · Released 21 Jul 2026 · gemini-3.5-flash-lite
$0.30
Input price / 1M tokens
#9 of 61, cheapest first
$2.50
Output price / 1M tokens
#14 of 61, cheapest first
1.05M
Context window
#11 of 71
360
Output tokens / second
#1 of 61
54.2%
SWE-bench Pro
#14 of 16
54%
Terminal-Bench 2.1
#17 of 17
74%
OSWorld-Verified
#7 of 8
74.5%
CharXiv Reasoning
#5 of 6
Summary
Gemini 3.5 Flash-Lite is a proprietary model from Google, released on 21 Jul 2026.
At $0.30 input and $2.50 output per million tokens, it is #9 of 61 on input price, cheapest first.
Its 1.05M-token context window ranks #11 of 71.
Measured output speed is 360 tokens per second, #1 of 61.
Low-latency, cost-efficient multimodal model for high-volume agentic sub-tasks, translation, document parsing and simple data processing.
Benchmarks · 7 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| SWE-bench Prohigh thinking, internal Antigravity harness | 54.2% | #14 of 16 | Google Gemini 3.5 Flash-Lite evaluation reportSelf-reported | |
| Terminal-Bench 2.1high thinking, Terminus 2 harness | 54% | #17 of 17 | Google Gemini 3.5 Flash-Lite evaluation reportSelf-reported | |
| MLE-BenchPartial-30 subset, average position score | 39.2% | – | Google Gemini 3.5 Flash-Lite evaluation reportSelf-reported | |
| GDPVal-AA v2Artificial Analysis leaderboard, Jul 2026 | 1,140 | Elo-style | – | Artificial Analysis, cited in Google's evaluation report |
| OSWorld-Verifiedaverage of 5 runs, max 100 steps | 74% | #7 of 8 | Google Gemini 3.5 Flash-Lite evaluation reportSelf-reported | |
| CharXiv Reasoningno tools (76.5% with search and code execution) | 74.5% | #5 of 6 | Google Gemini 3.5 Flash-Lite evaluation reportSelf-reported | |
| MRCR v2 (8-needle)128k average (21.3% at 1M pointwise) | 72.2% | #6 of 13 | Google Gemini 3.5 Flash-Lite 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 3.5 Flash-Lite, and the closest scores.
SWE-bench Pro
#14 of 16Harder multi-file engineering tasks
- Gemini 3.5 Flash-Lite54.2
- Gemini 3.5 Flash53.9
- Gemini 3.6 Flash58.7
- Qwen3.8-27B61.7
- GPT-5.457.7
Terminal-Bench 2.1
#17 of 17Terminal tasks, earlier set
OSWorld-Verified
#7 of 8Using a desktop computer
CharXiv Reasoning
#5 of 6Reading charts from research papers
Compare Gemini 3.5 Flash-Lite 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-3.5-flash-lite, 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-3-5-flash-lite/, checked 2 Oct 2026
- benchmarks: storage.googleapis.com/deepmind-media/gemini/gemini_3-5_flash-lite_model_evaluat, checked 2 Oct 2026
- output speed: artificialanalysis.ai/models/gemini-3-5-flash-lite, checked 2 Oct 2026
Building on Gemini 3.5 Flash-Lite?
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.