The AI model of the Mountain View-based tech giant, named Gemini 3.7 Flash, has been introduced globally to date on Thursday. This particular AI model supersedes the earlier one, known as Gemini 3.6 Flash. The earlier one came into play last month together with Gemini 3.5-Lite AI model. According to the company, the newer Gemini 3.7 model has enhanced abilities in software engineering, knowledge work, and web development workflows. The tech giant will be offering this AI model at the introductory pricing. It is also planning to update its Gemini Spark with the help of this new AI model.
According to Google, Gemini 3.7 Flash comes at half the introductory pricing of the older model, that is Gemini 3.6 Flash, which is priced at £1 per million tokens. Tech giant has already started releasing its newer Gemini 3 series AI model through Google Antigravity, Google AI Studio, and Android Studio platforms.
For enterprises, the Gemini 3.7 Flash model is available through the Gemini Enterprise Agent Platform and the Gemini Enterprise app. Individual Google AI Pro and Google Ultra subscribers can access this latest model through the Gemini app in select regions. The company is also upgrading its new personal AI agent, Gemini Spark, to Gemini 3.7 Flash, allowing users to access the new AI model from the same interface.
Google claims that Gemini 3.7 Flash offers improved performance in many respects compared to its previous version. This AI model is claimed to deliver superior intelligence for complex workflows, coding, workflow automation, document comprehension, web development, and long-horizon software engineering. On FrontierCode 1.1 Main, Gemini 3.7 Flash scored 43.6 percent, beating Gemini 3.6 Flash, Anthropic's Claude Sonnet 5, and OpenAI's GPT 5.6 Terra, while its rivals scored 34.4 percent, 42.7 percent, and 41.3 percent, respectively.
Similarly, on DeepSWE V1.1, Gemini 3.7 Flash scored 65.3 percent, beating Gemini 3.6 Flash's 48.6 percent, Claude Sonnet 5's 53.8 percent, and Meta's Muse Spark 1.2's 54.9 percent. However, GPT-5.6 Terra scored 69.6 percent, significantly higher than Gemini 3.7 Flash's score. In terms of web development, Gemini 3.7 Flash scored 1,588 points on Code Arena, while Gemini 3.6 Flash, Claude Sonnet 5, GPT-5.6 Terra, and Muse Spark 1.2 scored 1,538, 1,541, 1,523, and 1,535 points, respectively.
Google's new Gemini 3.7 Flash also outperformed its predecessor and competitors on the Expert PDF Document Comprehension test on GDP.pdf, scoring 34 percent. Compared to Gemini 3.6 Flash's 22 percent, Claude Sonnet 5's 28 percent, GPT-5.6 Terra's 24.7 percent, and Muse Spark 1.2's 16 percent. Similarly, for enterprise workflow automation, Gemini 3.7 Flash scored 30.4 percent on Automation Bench, significantly higher than Gemini 3.6 Flash, Cloud Sonnet 5, and GPT-5.6 Terra.
The AI model of the Mountain View-based tech giant, named Gemini 3.7 Flash, has been introduced globally to date on Thursday. This particular AI model supersedes the earlier one, known as Gemini 3.6 Flash. The earlier one came into play last month together with Gemini 3.5-Lite AI model. According to the company, the newer Gemini 3.7 model has enhanced abilities in software engineering, knowledge work, and web development workflows. The tech giant will be offering this AI model at the introductory pricing. It is also planning to update its Gemini Spark with the help of this new AI model.
According to Google, Gemini 3.7 Flash comes at half the introductory pricing of the older model, that is Gemini 3.6 Flash, which is priced at £1 per million tokens. Tech giant has already started releasing its newer Gemini 3 series AI model through Google Antigravity, Google AI Studio, and Android Studio platforms.
For enterprises, the Gemini 3.7 Flash model is available through the Gemini Enterprise Agent Platform and the Gemini Enterprise app. Individual Google AI Pro and Google Ultra subscribers can access this latest model through the Gemini app in select regions. The company is also upgrading its new personal AI agent, Gemini Spark, to Gemini 3.7 Flash, allowing users to access the new AI model from the same interface.
Google claims that Gemini 3.7 Flash offers improved performance in many respects compared to its previous version. This AI model is claimed to deliver superior intelligence for complex workflows, coding, workflow automation, document comprehension, web development, and long-horizon software engineering. On FrontierCode 1.1 Main, Gemini 3.7 Flash scored 43.6 percent, beating Gemini 3.6 Flash, Anthropic's Claude Sonnet 5, and OpenAI's GPT 5.6 Terra, while its rivals scored 34.4 percent, 42.7 percent, and 41.3 percent, respectively.
Similarly, on DeepSWE V1.1, Gemini 3.7 Flash scored 65.3 percent, beating Gemini 3.6 Flash's 48.6 percent, Claude Sonnet 5's 53.8 percent, and Meta's Muse Spark 1.2's 54.9 percent. However, GPT-5.6 Terra scored 69.6 percent, significantly higher than Gemini 3.7 Flash's score. In terms of web development, Gemini 3.7 Flash scored 1,588 points on Code Arena, while Gemini 3.6 Flash, Claude Sonnet 5, GPT-5.6 Terra, and Muse Spark 1.2 scored 1,538, 1,541, 1,523, and 1,535 points, respectively.
Google's new Gemini 3.7 Flash also outperformed its predecessor and competitors on the Expert PDF Document Comprehension test on GDP.pdf, scoring 34 percent. Compared to Gemini 3.6 Flash's 22 percent, Claude Sonnet 5's 28 percent, GPT-5.6 Terra's 24.7 percent, and Muse Spark 1.2's 16 percent. Similarly, for enterprise workflow automation, Gemini 3.7 Flash scored 30.4 percent on Automation Bench, significantly higher than Gemini 3.6 Flash, Cloud Sonnet 5, and GPT-5.6 Terra.