Google Unleashes Gemini 3 Flash, Challenging OpenAI with Faster, Cheaper AI

Google has launched its new Gemini 3 Flash AI model. The announcement was made today. The model is now the default in the Gemini app and AI search.

This fast and affordable model aims to compete directly with OpenAI. It is based on the more powerful Gemini 3 Pro architecture released last month. According to TechCrunch, the rollout is happening globally.

Benchmark Performance Shows Major Leap

The new model shows significant performance gains. It far outpaces its predecessor, Gemini 2.5 Flash, on key tests. It also matches frontier models in some areas.

For example, it scored 33.7% on the Humanity’s Last Exam benchmark. Gemini 3 Pro scored 37.5% on the same test. The new GPT-5.2 model from OpenAI scored 34.5%.

On the MMMU-Pro multimodal reasoning benchmark, it led all rivals. It achieved a top score of 81.2%. This demonstrates strong capabilities in understanding mixed content like images and text.

Consumer and Enterprise Rollout Begins

For everyday users, Gemini 3 Flash is now the default in the Gemini app. It replaces the older Flash model. Users can still select the Pro model for complex tasks like coding.

The model excels at multimodal understanding. You can upload a short video for sports tips. You can also draw a sketch for it to identify or analyze an audio recording.

For businesses, the model is available now. Companies like JetBrains, Figma, and Cursor are already using it. Access is provided through Google’s Vertex AI and Gemini Enterprise platforms.

Developers can try it via API preview. It is also in Antigravity, Google’s coding tool. The company says it processes over a trillion tokens daily on its APIs.

Model pricing is set at $0.50 per million input tokens. Output tokens cost $3.00 per million. Google claims it outperforms the older Pro model while being three times faster.

The release of Gemini 3 Flash intensifies the AI arms race with OpenAI. Google’s bet on a faster, cheaper “workhorse” model could reshape how businesses and consumers use artificial intelligence. This move directly challenges the current market dynamics.

A quick knowledge drop for you

What is Google Gemini 3 Flash best for?

It is designed as a fast, cost-effective “workhorse” for bulk tasks. This includes quick visual Q&A, data extraction, and video analysis. Its speed makes it ideal for repeatable workflows.

How much does Gemini 3 Flash cost?

Pricing is $0.50 for every 1 million input tokens. Generating 1 million output tokens costs $3.00. This is slightly more than the previous Flash model but offers much higher performance.

Can I use Gemini 3 Flash for free?

Yes, it is the new default model in the free Gemini app. Users can access it immediately for general queries. The Pro model remains available in the app’s model picker for specialized needs.

How does it compare to OpenAI’s GPT-5.2?

Benchmarks show very close performance on general knowledge tests. Gemini 3 Flash scored 33.7% versus GPT-5.2’s 34.5% on one key exam. Google’s model took the lead on the multimodal MMMU-Pro benchmark.

Why is Google releasing this model now?

The release follows intense competition with OpenAI. Reports indicated a “Code Red” at OpenAI after Google gained consumer market share. This launch is a direct counter to OpenAI’s recent GPT-5.2 update.