Google is stepping up its AI competition with the launch of Gemini 3.8 Flash, which it positions as its most capable model yet for coding. The company says the new model can deliver performance comparable to leading systems such as Claude Opus 5 and GPT-5.6 Sol while costing significantly less to operate. Google CEO Sundar Pichai announced the release, marking the company’s third Flash model launch in just six weeks. As a smaller, more efficient model, Gemini Flash is designed to offer strong performance with lower computing and running costs.
Gemini 3.8 Flash Claims Flagship-Level Coding at Lower Cost #
Google says Gemini 3.8 Flash can tackle complex engineering and coding problems with performance that surpasses many larger AI models, while requiring a fraction of the cost. The company also highlights improvements over Gemini 3.7 Flash in software development and multi-step reasoning. The model is initially priced at $0.75 per million input tokens and $3.75 per million output tokens through December 31, 2026, before those rates double. However, its ability to spend more tokens on difficult tasks could increase the overall bill for some users.
Today we’re introducing a new 3.8 Flash model, our 3rd Flash release in just 6 wks.
— Sundar Pichai (@sundarpichai) It delivers significant leaps from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning. On DeepSWE v1.1, it outperforms most larger frontier models in autonomously…[pic.twitter.com/MB5seliluh][September 2, 2026]
How Does Gemini 3.8 Flash Compare on Price? #
The pricing could become one of Gemini 3.8 Flash’s biggest selling points, particularly for businesses using AI for coding and other demanding workloads. Even after the introductory period, Google’s rates remain below those listed for Claude Opus 5 and GPT-5.6 Sol. The company is betting that this combination of capability and lower per-token pricing can help it compete more aggressively in the enterprise AI market. At the same time, cheaper open-weight alternatives such as Kimi K3 are putting additional pressure on major AI companies to control costs.
Google Says Gemini 3.8 Flash Is a Major Reasoning Upgrade #
Gemini 3.8 Flash scored 54.9 percent on HLE-Verified, according to Google, reflecting its ability to work through complex, multi-stage problems across several fields. A key change is the model’s ability to increase its reasoning effort when faced with harder tasks, including repeatedly using tools to reach an answer. That extra work can improve results but also consume more tokens. Artificial Analysis reportedly found that usage rose by around 40 percent compared with Gemini 3.7 Flash during agentic evaluations, making the advertised token price only part of the cost equation.
Gemini 3.8 Flash Cyber Targets the AI Security Race #
Google has also introduced Gemini 3.8 Flash Cyber, a specialised model aimed at cybersecurity defenders and vulnerability research. The company says it generated 2.6 times more correct patches for Chrome vulnerabilities than the best larger commercial models tested by its security team, while another Google team used it to identify a critical vulnerability in under two hours. The launches underline Google’s effort to regain momentum against OpenAI and Anthropic in coding and specialised AI, as the company also works toward its next flagship model, Gemini 4.
ALSO SEE:
[Can Gemini Help MrBeast Survive the Arctic? Google Puts AI to the Ultimate Test](/tech/113543/can-gemini-help-mrbeast-survive-the-arctic-google-puts-ai-to-the-ultimate-test)[Tech](/tech),
[Sundar Pichai](/sundar-pichai),
[ai](/ai),
[Openai](/openai),
[Gemini](/gemini),
[ChatGPT](/chatgpt),
[Perplexity](/perplexity),
[Anthropic](/anthropic),
[Gemini AI](/gemini-ai),
[Gemini 3.8 Flash](/gemini-38-flash)