gemini-2.5-pro-preview-03-25 (Initial 2.5 Pro Experimental)
Released on March 25, 2025, this was the first public build of Gemini 2.5 Pro. It introduced the “thinking model” architecture—meaning the model reasons through chain-of-thought steps internally before generating its output—and shipped with a 1 million-token context window. At launch, it set new SOTA marks on reasoning and STEM benchmarks (e.g. 18.8 % on Humanity’s Last Exam, AIME 2025 pass@1 of 86.7 %) and demonstrated advanced code-generation/editing capabilities (scoring 63.8 % on SWE-Bench Verified) without requiring ensemble or majority-voting tricks.
gemini-2.5-pro-preview-05-06 (I/O Edition)
Rolled out on May 6, 2025, just ahead of Google I/O, this “I/O Edition” of 2.5 Pro (internally labeled gemini-2.5-pro-preview-05-06) focused heavily on improving programming performance. Compared to the March 25 build, it delivers major upgrades in code transformation, code editing, and support for complex, agentic workflows—making it noticeably better at generating and refactoring production-quality software. It also continued to lead top human-preference and academic benchmarks (e.g. LMArena, AIME 2025, GPQA Diamond) without test-time hacks.
gemini-2.5-pro-preview-06-05 (Post-I/O Update)
Deployed on June 5, 2025, this build added several new “big-picture” features beyond the I/O Edition optimizations. Namely, it introduced Deep Think mode—an explicit toggle for deeper chain-of-thought reasoning—as well as native audio-output support and enhanced security controls. These additions further bolster Gemini 2.5 Pro’s ability to tackle complex, multimodal tasks (text, code, audio, video) with more reliable, context-aware outputs. The model still uses a 1 million-token window (2 million tokens coming soon) but now offers the Deep Think reasoning switch for even more thorough internal deliberation .
Gemini 2.5 Pro API
How to call \**Gemini 2.5 pro\** API from CometAPI
\**Gemini 2.5 pro\** API Pricing in CometAPI，20% off the official price:
Input Tokens: $1/ M tokens
Output Tokens: $8/ M tokens
Required Steps
Log in to cometapi.com. If you are not our user yet, please register first
Get the access credential API key of the interface. Click “Add Token” at the API token in the personal center, get the token key: sk-xxxxx and submit.
Select the “g\**emini-2.5-pro-preview-06-05\**” endpoint to send the API request and set the request body. The request method and request body are obtained from our website API doc. Our website also provides Apifox test for your convenience.
Replace <YOUR_AIMLAPI_KEY> with your actual CometAPI key from your account.
Insert your question or request into the content field—this is what the model will respond to.
. Process the API response to get the generated answer.
Gemini 2.5 Pro stands as a testament to the evolving nature of AI technology. With its advanced reasoning capabilities, multi-modal input support, and robust application scenarios, it heralds a new era for developers and users alike. As this model continues to evolve, it promises to unlock unprecedented opportunities across diverse fields, reinforcing Google’s position as a leader in artificial intelligence development.
