Gemini 3.8 Flash
Google · Released 2 Sep 2026 · gemini-3.8-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
249
Output tokens / second
#5 of 61
73.7%
DeepSWE v1.1
#5 of 17
89.4%
Terminal-Bench 2.1
#2 of 17
19.1%
Terminal-Bench 4.0
#9 of 10
59%
OSWorld 2.0
#7 of 10
Summary
Gemini 3.8 Flash is a proprietary model from Google, released on 2 Sep 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 249 tokens per second, #5 of 61.
Google's most intelligent Flash model, built for long-horizon software engineering, autonomous agents and complex enterprise workflows at Flash speed and cost.
Benchmarks · 13 reported
| Benchmark | Score | Bar | Rank | Source |
|---|---|---|---|---|
| DeepSWE v1.1mini-swe agent harness, high thinking | 73.7% | #5 of 17 | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| Terminal-Bench 2.1Terminus 2 harness | 89.4% | #2 of 17 | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| Terminal-Bench 4.0from official public leaderboard, highest-scoring thinking level | 19.1% | #9 of 10 | Terminal-Bench leaderboard, cited in Google's evaluation report | |
| HLE-Verifiedfull 1,811-item verified set | 54.9% | – | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| OSWorld 2.0partial score, batched tool calls, max over 3 runs, single attempt per run | 59% | #7 of 10 | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| GDPVal-AA v2Artificial Analysis leaderboard | 1,545 | Elo-style | – | Artificial Analysis, cited in Google's evaluation report |
| Vals Finance Agent v2 | 61.4% | – | Vals.ai, cited in Google's evaluation report | |
| Harvey's Legal Agent Benchmarkall-pass rate | 10% | – | Vals.ai, cited in Google's evaluation report | |
| GDP.PDFall-pass rate | 35% | – | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| CharXiv Reasoningno tools | 86.2% | #1 of 6 | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| LVBenchstatic, no tools, 1024 frames (87.8% in agentic mode) | 87.1% | – | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| BioMysteryBenchhuman-solvable set (56.5% on human-difficult set) | 88.8% | – | Google Gemini 3.8 Flash evaluation reportSelf-reported | |
| LABBench2terminal with bioinformatics tools and internet; macro-average of 11 sub-tasks | 86.2% | – | Google Gemini 3.8 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 3.8 Flash, and the closest scores.
DeepSWE v1.1
#5 of 17Long coding tasks in real repos
- Gemini 3.8 Flash73.7
- Gemini 3.7 Flash65.3
- GPT-5.6 Sol72.7
- GPT-5.6 Terra69.6
- GPT-5.6 Luna67.2
- Gemini 3.6 Flash49
Terminal-Bench 2.1
#2 of 17Terminal tasks, earlier set
- Gemini 3.8 Flash89.4
- Gemini 3.7 Flash85.8
- GPT-5.6 Sol88.8
- GPT-5.6 Terra87.4
- GPT-5.6 Luna84.7
- Gemini 3.6 Flash78
OSWorld 2.0
#7 of 10Long desktop and web workflows
- Gemini 3.8 Flash59
- Gemini 3.7 Flash47.9
- GPT-5.6 Sol62.6
- GPT-5.6 Terra50.2
- GPT-5.6 Luna45.6
Compare Gemini 3.8 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-3.8-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-3-8-flash/, checked 2 Oct 2026
- benchmarks: storage.googleapis.com/deepmind-media/gemini/gemini_3-8_flash_model_evaluation.p, checked 2 Oct 2026
- launch post: blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-, checked 2 Oct 2026
- output speed: artificialanalysis.ai/models/gemini-3-8-flash, checked 2 Oct 2026
Building on Gemini 3.8 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.