Google's Gemini 3 takes the crown

PLUS: OpenAI's new coding agent and Google's Nano Banana Pro arrives

Author

Gen AI Art
November 21, 2025 • Estimated Reading Time: 8 minutes

Together with

Google has just unveiled its latest foundational model, Gemini 3, and it has already claimed the top spot on a key industry leaderboard. The new model sets a new benchmark for performance across reasoning, math, and multimodal understanding.

With record-breaking scores on graduate-level science questions and complex problem-solving, the release intensifies the competition among top AI labs. Does this leap in capability signal a new frontrunner in the race for AI dominance?

Today in AI:

  • Google's Gemini 3 takes the crown

  • OpenAI's new coding agent

  • The emerging AI platform strategy

PRESENTED BY BRIGHT DATA

74% of Companies Are Seeing ROI from AI.

Incomplete data wastes time and stalls ROI. Bright Data connects your AI to real-time public web data so you launch faster, make confident decisions, and achieve real business growth.

Google's Gemini 3 Takes The Crown

What’s new? Google has launched Gemini 3, its new foundational model that has already secured the top spot on the LMArena Leaderboard. It sets a new benchmark for performance in reasoning, math, and multimodal understanding.

What matters?

  • Gemini 3 demonstrated powerful performance on diverse expert-level tasks, scoring 91.9% on GPQA Diamond for graduate-level science questions.

  • The model established a new record on advanced mathematical problem-solving and scored 37.5% on Humanity's Last Exam, a test designed for PhD-level reasoning.

  • It also achieved state-of-the-art results on SpatialBench, a benchmark testing an AI’s ability to comprehend 3D relationships and object positioning in images.

Why it matters?

Google's latest release intensifies the competition among top AI labs, proving that leadership in model performance is still very much contested. This leap in reasoning and multimodal understanding paves the way for more capable and reliable AI applications across professional and creative fields.

OpenAI's Counter-Strike for Coders

What’s new? In a direct shot at competitors, OpenAI has released GPT-5.1-Codex-Max, a new frontier coding model designed to reclaim the top spot for AI-assisted software development.

What matters?

  • The model introduces a compaction feature, allowing it to work on large-scale projects for over 24 hours by automatically pruning its session history to retain only relevant context.

  • It achieved top scores on the SWE-Bench Verified benchmark with 77.9% accuracy, putting it ahead of both Anthropic's and Google’s latest models for fixing real-world code bugs.

  • The new model is available now for Plus, Pro, Business, and Enterprise users within Codex.

Why it matters?

This move signals OpenAI's intense focus on winning over developers, who are the primary drivers of AI adoption and token consumption. The race for AI leadership is increasingly being fought by creating the most capable and indispensable tools for the people who build with them.

PRESENTED BY MORNING BREW

Trusted by millions. Actually enjoyed by them too.

Morning Brew makes business news something you’ll actually look forward to — which is why over 4 million people read it every day.

Sure, the Brew’s take on the news is witty and sharp. But the games? Addictive. You might come for the crosswords and quizzes, but you’ll leave knowing the stories shaping your career and life.

Try Morning Brew’s newsletter for free — and join millions who keep up with the news because they want to, not because they have to.

The New AI Platform Strategy

What’s new? A new platform strategy is emerging where companies like Adobe and Microsoft integrate various top-tier AI models—from Google, OpenAI, and others—directly into their existing applications, treating them as swappable commodities.

What matters?

  • Instead of forcing users to chase new model releases across different sites, this approach brings the best tools directly into established workflows, a finding supported by Adobe's recent study showing most creators already use multiple AI tools.

  • Adobe is a prime example, recently adding Google's Nano Banana Pro to Firefly and Photoshop, which now sit alongside models from OpenAI, Runway, Pika, and others.

  • Microsoft is pursuing a similar path, integrating models like Anthropic's Claude into its Azure and Copilot ecosystems to offer customers more choice.

Why it matters?

This strategy shifts the battleground from who builds the best single model to which platform provides the most seamless access to all of them. For users, it means the best AI capabilities become readily available features in the software they already use daily.

Everything else in AI

OpenAI detailed the capabilities, limitations, and safety measures for its new GPT-5.1-Codex-Max model in a comprehensive system card.

Adobe found in a recent study that over 60% of creators already use multiple AI models, reinforcing the industry shift toward integrated AI platforms.

Reddit highlighted the impressive photorealism of Google's new Nano Banana Pro model, with users struggling to distinguish its AI-generated images from real photographs.

Essential AI Guides - Reading List:

Let us know!

Work with us

Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now →