Following last month’s launch of Gemini 3 Pro, Google today announced Gemini 3 Flash for consumers and developers.
The tagline, of sorts, for Gemini 3 Flash is “frontier intelligence built for speed at a fraction of the cost.” It retains Gemini 3’s complex reasoning, multimodal/vision understanding, and performance in agentic/vibe coding tasks, but adds “Flash-level latency, efficiency and cost.” The Flash model series is Google’s most popular offering
It unsurprisingly surpasses 2.5 Flash across the board and “significantly” outperforms 2.5 Pro across several benchmarks. 3 Flash is comparable to 3 Pro, while even beating it in some areas, including MMMU Pro, Toolathlon, and MPC Atlas.
- GPQA Diamond (scientific knowledge): 90.4%
- Humanity’s Last Exam (academic reasoning): 33.7% without tools
- MMMU Pro (multimodal understanding/reasoning): 81.2%
- SWE-Bench Verified (agentic coding): 78%
- Toolathlon (long horizon real-world software tasks): 49.4%
- MCP Atlas (multi-step workflows using MCP): 57.4%
Google touts how 3 Flash “outperforms 2.5 Pro while being 3x faster at a fraction of the cost.” This strong reasoning, tool use, and multimodal capabilities can translate to “more complex video analysis, data extraction and visual Q&A” for third-party developers building customer support agents or in-game assistants.
