Gemini 3 Flash Preview
google/gemini-3-flash-preview- Family · Gemini 3 Flash
- Released 17 Dec 2025
- Text · Image · File · Audio · Video → Text
- Reasoning · adjustable
- Tools · JSON · Vision
Spec & list price
Context1M
Max output64K
Input / 1M$0.50
Output / 1M$3.006× input
Scores · standing
Intelligence26.3#85/489
Usage
Tokens / day88.1B▼ 15% 7d
Usage rank#34
Requests / day7.0M
Tool calls / req0.09
Per request · 30 days
Avg input9.4K0.6× median
Avg output3720.5× median
Cache hit39%0.6× median
Reasoning share20%0.4× median
Tool-call failures2.6%1.4× median
Consumption
Peak 190.7B 12 OctLow 66.7B 1 NovRank today #34
News & trends
1 mentions this week ▼ 83%this model · videos, posts, articles · 30 days
Trending · 14 days
- GLM 5.1 ranked between GLM 5.3 and GLM 5. Where would you rank Gemini 3 Flash Preview?13 upvotes
- Gemini 3 Flash, “el loco”: guerra nuclear simulada en 4 turnos #Shorts45 views
- GPT-5.2 và Gemini 3 Flash Thống Trị Bảng Xếp Hạng Toán Học AI16 views
News · latest from Gemini 3 Flash
- Claude Opus 5.5 and Claude Sonnet 5.5 are now available in Google AntigravityGoogle Cloud · 20 Oct
- Introducing Gemini 3.8 Flash and 3.8 Flash CyberGoogle DeepMind · 2 Sep
- Introducing Gemini 3.7 FlashGoogle DeepMind · 13 Aug
- Introducing Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash CyberGoogle DeepMind · 21 Jul
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across 6 providers, latest day
Output speed#97/33491 tok/s1.4× faster than median
Fastest provider146 tok/sbest of 6 providers
First token#2221.52 s1.5× longer than median
#12 cheapest of 25 with a similar score
Value
Intelligence Index
Blended price per 1M tokens (log)
Family progress
Intelligence Index
- Family releases40.9
Technical details
- Reasoning
- on · medium
- Tokenizer
- Gemini
- Moderation
- none
Reasoning effort
- high
- medium · default
- low
- minimal
- none
Request parameters 12
- tools
- tool_choice
- response_format
- structured_outputs
- reasoning
- reasoning_effort
- include_reasoning
- temperature
- top_p
- max_tokens
+2 more
- stop
- seed
Interesting facts
- Less output room than similar models (65.5K against a 309.7K median).
- 2nd of 5 Gemini 3 Flash models by tokens last month.
When to use
Best for
- Fast coding and math assistance
- Instruction-following in chat apps
- Video, audio and image understanding
- Long document Q&A
Think twice for
- Production use that needs stable behavior (preview)
- Self-hosting (closed weights)
- Image generation
Research papers
- Not Where You Think: Protecting Location Privacy Against Reasoning-Based Geolocation21 Oct
- Combining Long-Term Memory and LLMs Comes at a High Cost: From Prompt Injection to Persistent Compromise20 Oct
- AM-Bench: A Unified Taxonomy and Evaluation Suite for Agentic Misalignment16 Oct
- BRQGen: Towards Adaptive Completeness Assessment and Follow-Up Question Generation for Bug Reports11 Oct
- VideoResearchAgent: Grounded Task Synthesis and Sim-to-Real RL for Open-Web Video Research4 Oct
- AgentWorld: Benchmarking Long-Horizon Collaboration of Multi-agent LLMs25 Sep
- Prefilling the Reasoning Channel: Output-Prefix Attacks on Reasoning LLMs24 Sep
- System Attribution in LLM Brand Recommendations: Single Responses Identify the System, Aggregated Brand Profiles Do Not Transfer23 Sep
Social
Most watched and discussed · 90 days
- Gemini 3 Flash, “el loco”: guerra nuclear simulada en 4 turnos #ShortsSTRATEGA · 45 views · 21 Oct
- GPT-5.2 và Gemini 3 Flash Thống Trị Bảng Xếp Hạng Toán Học AIAI Race - Cuộc Đua AI · 16 views · 20 Oct
- Gemini 4 Argon VS Gemini 3 FlashRam Kumar Sharma · 8 views · 27 Oct
- #32 Gemini 3 Flash, New Claude Code Features, Framer Awards, Github Prices, GPT ImagesCmd+AI · 1 views · 20 Oct
Same family
Gemini 3 Flash- 40.9
- 33.6
- 34.0
- 39.6
Alternatives
other vendors · similar capability- 29.2
- 21.8
- T25.0
- 21.5
Using Gemini 3 Flash Preview at volume?
Sign in to ask for a volume discount. You will continue in the OSMEXO dashboard.