Google Gemini 2.5

Key Takeaways

• Advanced Reasoning: Scores 68.6% on Aider Polyglot and 63.8% on SWE-Bench verified in coding tasks.
• Larger Context: Supports 1M tokens, expanding to 2M, exceeding 750,000 words.
• Pricing & Availability: Priced at $20/month, with Vertex AI integration coming soon.

Google Unveils Gemini 2.5 with Advanced Reasoning Capabilities

On Tuesday, Google (GOOGL) introduced Gemini 2.5, its most advanced AI model yet, featuring enhanced reasoning capabilities that allow it to “think” before responding. This multimodal AI model is designed for complex problem-solving, outperforming previous Gemini models and competing AI systems on multiple benchmarks. Notably, Gemini 2.5 Pro scores 68.6% on Aider Polyglot for code editing, outperforming top models from OpenAI, Anthropic, and DeepSeek. 

On SWE-Bench Verified, it achieves 63.8%, surpassing OpenAI’s o3-mini and DeepSeek’s R1 but falling short of Anthropic’s Claude 3.7 Sonnet at 70.3%. Additionally, on Humanity’s Last Exam, which evaluates multimodal reasoning across thousands of crowdsourced questions in mathematics, humanities, and natural sciences, it scores 18.8%, ranking above most flagship AI models.

Extended Context Window and Performance Enhancements

One of the model’s standout features is its ability to process extensive amounts of context in a single go, starting with a 1 million token context window—equivalent to about 750,000 words, longer than the entire Lord of the Rings book series. Google plans to double this limit to 2 million tokens, further boosting its ability to handle large-scale documents and complex tasks. 

Gemini 2.5 Pro has also topped the LMArena leaderboard, which measures human preferences in AI responses and has demonstrated strong performance in AIME 2025 (math) and GPQA diamond (science) benchmarks without relying on expensive test-time techniques like majority voting. Its coding capabilities have significantly improved over previous versions, making it a powerful tool for developing visually compelling web apps and agentic coding applications.

Availability and Future Integrations

Gemini 2.5 Pro is now available in Google AI Studio and the Gemini app for $20/month Gemini Advanced subscribers, with upcoming integration into Vertex AI, enabling developers and enterprises to scale AI-powered solutions with higher rate limits and enterprise-grade performance.

Google has confirmed that all future AI models will include built-in reasoning capabilities, ensuring they are more context-aware, accurate, and effective in solving complex tasks. API pricing for Gemini 2.5 Pro is expected to be announced soon, offering scalable solutions for businesses and developers.

As Google continues refining Gemini’s capabilities, it actively seeks user feedback to further enhance its multimodal intelligence, advanced reasoning, and coding proficiency.