Google Launches Gemini 3 Pro AI: A Rival to ChatGPT, Grok, and Claude
Google has officially launched Gemini 3 Pro, positioning it as a cutting-edge entrant in the competitive AI landscape. This follows the previous success of the Gemini 2.5 Pro model, which was once considered a top performer until overtaken by Elon Musk’s Grok AI in certain tests.
Gemini 3 Pro: Setting New Benchmarks
The new Gemini 3 Pro model not only surpasses its predecessor but also outperforms major competitors such as ChatGPT and Claude according to benchmarks released by Google. On the LMArena leaderboard, Gemini 3 Pro achieved an impressive score of 1501 in text-related tasks, making it the top model ahead of Grok 4.1.
- LMArena Leaderboard: Gemini 3 Pro ranks first with a score of 1501.
- Coding Excellence: The model also tops coding, math, and creative writing tasks across various leaderboards.
Academic and Mathematical Performance
In the field of academic reasoning, the Gemini 3 Pro excelled on Humanity’s Last Exam, scoring 37.5%. This puts it significantly ahead of GPT-5.1, which scored 26.5%, and Claude Sonnet 4.5, trailing at 13.7%.
- Humanity’s Last Exam: Gemini 3 Pro: 37.5%, GPT-5.1: 26.5%, Claude Sonnet 4.5: 13.7%.
Gemini 3 Pro also demonstrated superior capabilities in mathematical challenges on the MathArena Apex benchmark, achieving a score of 23.4%. In comparison, Gemini 2.5 Pro and other models like Claude Sonnet 4.5 and GPT-5.1 scored between 0.5% and 1.6%.
Performance in Screen Understanding
On the ScreenSpot Pro benchmark, designed to test a model’s understanding of computer screens, Gemini 3 Pro scored 72.7%. This score significantly outperformed both Claude Sonnet 4.5 and GPT-5.1, scoring 36.2% and 3.5%, respectively.
Challenges Remain in Coding Tasks
Despite its successes, Gemini 3 Pro faced challenges in specific coding benchmarks. For example, on the SWE-Bench Verified test, Claude Sonnet 4.5 attained the highest score of 77.2%, while Gemini 3 Pro secured third place with 76.2%.
- SWE-Bench Verified:
- Claude Sonnet 4.5: 77.2%
- GPT-5.1: 76.3%
- Gemini 3 Pro: 76.2%
Future of Gemini 3 Pro
As AI development accelerates, the lead held by Gemini 3 Pro may be temporary, given the frequent release of new models by various companies. Nonetheless, it currently stands out as a leading AI model across many benchmarks.
It is important to recognize that benchmark scores may not fully encapsulate an AI model’s capabilities. Actual performance is often best assessed through real user experiences.
