Google's Gemini 2.5 Pro: Advanced AI Model Now Available for Free with Limited Access

Google's new experimental Gemini 2.5 model rolls out to free users

Google released its latest and greatest Gemini AI model last week, but it was only made available to paying subscribers. Google has moved with uncharacteristic speed to release Gemini 2.5 Pro (Experimental) for free users, too. The next time you check in with Gemini, you can access most of the new AI's features without a Gemini Advanced subscription. The Gemini 2.5 branch will eventually replace 2.0, which was only released in late 2024. It supports simulated reasoning, as all Google's models will in the future. This approach to producing an output can avoid some of the common mistakes that AI models have made in the past. We've also been impressed with Gemini 2.5's vibe, which has landed it at the top of the LMSYS Chatbot arena leaderboard. Google says Gemini 2.5 Pro (Experimental) is ready and waiting for free users to try on the web. Simply select the model from the drop-down menu and enter your prompt to watch the "thinking" happen. The model will roll out to the mobile app for free users soon. While the free tier gets access to this model, it won't have all the advanced features. You still cannot upload files to Gemini without a paid account, which may make it hard to take advantage of the model's large context window -- although you won't get the full 1 million-token window anyway. Google says the free version of Gemini 2.5 Pro (Experimental) will have a lower limit, which it has not specified. We've added a few thousand words without issue, but there's another roadblock in the way.

Gemini 2.5 Pro is Google's most expensive AI model yet | TechCrunch

On Friday, Google released API pricing for Gemini 2.5 Pro, an AI reasoning model with industry-leading performance on several benchmarks measuring coding, reasoning, and math. For prompts up to 200,000 tokens, Gemini 2.5 Pro costs $1.25 per million input tokens (roughly 750,000 words, longer than the entire "Lord of The Rings" series) and $10 per million output tokens. For prompts greater than 200,000 tokens (which most of Google's competitors don't support), Gemini 2.5 Pro costs $2.50 per million input tokens, and $15 per million output tokens. That pricing makes Gemini 2.5 Pro more expensive for developers than any other AI model currently offered by Google, including Gemini 2.0 Flash ($0.10/M input tokens, $0.40/M output tokens). It also makes Gemini 2.5 Pro more expensive than several other frontier AI models, such as OpenAI's o3-mini ($1.10/M input tokens, $4.40/M output tokens) and DeepSeek's R1 ($0.55/M input tokens, $2.19/M output tokens). To be fair, Gemini 2.5 Pro, which is available for free with strict rate limits, comes in cheaper than some other highly competitive models, including Anthropic's Claude 3.7 Sonnet ($3/M input tokens, $15/M output tokens) and OpenAI's GPT-4.5 ($75/M input tokens, $150/M output tokens). The tech industry's initial reaction has been largely positive, with developers praising what they perceive to be sensible rates. But broadly speaking, there seems to be some upward pressure on pricing for flagship models. The cost of recent top-of-the-line releases from labs like Google, OpenAI, and Anthropic has been going up, not down. See, for example, OpenAI's recently launched o1-pro, which is the company's most expensive API offering yet at $150/M input tokens and $600/M output tokens. It could be that high demand and computing costs are driving the trend. According to Google CEO Sundar Pichai, Gemini 2.5 Pro is the company's most in-demand AI model among developers, leading to an 80% increase in usage in Google's AI Studio platform and the Gemini API this month alone.

Everyone can now try Gemini 2.5 Pro - for free

Moments after DeepSeek released its latest model, another AI giant has already stolen back some of the limelight. Last Tuesday, Google announced Gemini 2.5, its "most intelligent" model. The company said this initial release is an "experimental version of 2.5 Pro, which is state-of-the-art on a wide range of benchmarks and debuts at #1 on LMArena by a significant margin." Then, on Saturday, Google revealed 2.5 Pro is now available to all Gemini users at gemini.google.com. Also: You can access free Gemini Gems on Android and iOS now - where to find them A family of thinking models, meaning they reason through their responses, Google's Gemini 2.5 release follows the Gemini 2.0 Flash Thinking models that landed in December. Gemini 2.5 Pro Experimental outperforms OpenAI's o3 mini and Anthropic's Claude 3.7 Sonnet on Humanity's Last Exam (HLE), a recently created benchmark designed to combat saturation -- or the problem of industry tests becoming too easy for rapidly evolving models. HLE is, therefore, a relatively harder test to perform well on. Gemini 2.5 scored 18.8% compared to o3 mini's 14% (evaluated using text problems only, no images) and Claude 3.7 Sonnet's 8.9%. Also: Worried about DeepSeek? Turns out, Gemini is the biggest data offender Already topping the Chatbot Arena leaderboard, the new model also outperformed competitors on common benchmarks for science, math, and coding, though usually by a smaller margin, which is now expected given the rate at which new models are accelerating. Google reported that Gemini 2.5 Pro Experimental shows improvements in reasoning, multimodal, and agentic capabilities, even from a "single line prompt." The new model also scored higher than its competitors on an IQ test hosted by testing site Tracking AI, which uses bespoke questions that are not publicly available and thus can't be included in training data. However, experts caution that human IQ tests -- beyond being questionable for having roots in eugenics -- are not exactly a useful measure of an AI model's capabilities, because human modalities of intelligence operate in significantly different ways. Google posted on X that Gemini 2.5 Pro is available to all Gemini users "with rate limits," after an initially more narrow release, and it will be coming to mobile soon. While the company did not update the specifics, it reiterated that Gemini Advanced users still have "expanded access" in addition to a larger context window.

Google's new 'reasoning' Gemini AI model is now available for free.

Less than a week after its release to Gemini Advanced subscribers, the Google Gemini 2.5 Pro (experimental) model is now available to anyone who wants to try it, albeit with rate limits. Google says the new models now incorporate step-by-step "thinking" processes to become the best ones yet, and outpace the competition on LLM benchmarks. Google says it's also working on adding the new model to its mobile app, while Advanced subscribers still have access to a larger context window, and they can now use the new model in Gemini's Canvas tool for live coding projects.

Google's new experimental AI model, Gemini 2.5 Pro, is now available to free users too

The company introduced the new model last week, first bringing it to Advanced users. Non-paying Gemini users can now play around with Google's newest model, the experimental version of Gemini 2.5 Pro. The company this weekend that it's making Gemini 2.5 Pro (experimental) free for everyone to use, albeit with tighter rate limits for non-subscribers. Google , touting it as its "most intelligent AI model" yet, and rolled it out to Gemini Advanced users first. It's available now in Google AI Studio and the Gemini app. While free users can now try it out too, Google added that "Gemini Advanced users have expanded access and a significantly larger context window." Gemini 2.5 Pro (experimental) is the first of Google's Gemini 2.5 "thinking" models, which are said to deliver more accurate results through reasoning. In a , the company explained that this "refers to its ability to analyze information, draw logical conclusions, incorporate context and nuance, and make informed decisions."

Google is now letting free users try out Gemini 2.5 Pro

Google has touted this model as being its most intelligent model to date. In a recently published blog, the company claims it outperforms some of the most popular AI models across a range of benchmarks, like science and mathematics. However, Gemini 2.5 Pro (experimental) isn't the only thing Google is bringing to a wider audience. Earlier this month, Google showed off Gemini Live's new live video capabilities to attendees at MWC 2025. This feature allows Gemini to take over your phone's camera to see what you see and provide answers to your questions. The feature has since rolled out to Gemini Advanced subscribers, but Google has revealed in a social media post that it plans to make the feature available to more users. Google did not specify when the feature will be released more widely.

Google Gemini's powerful new 2.5 Pro model is already available to free users

Summary Google Gemini has rolled out its powerful new 2.5 Pro experimental model to non-paying users, just days after its arrival on Gemini Advanced. Despite being available to free users, there will be some restrictions in place, such as rate limits. The Gemini 2.5 Pro experimental model brings improvements to coding, math, and science queries, while also outranking industry competitors in benchmarks. Google Gemini recently unleashed its latest and greatest 2.5 Pro model, calling it their "most intelligent AI model." Much like most model updates or launches, Google said it will be available exclusively to subscribers of Gemini Advanced. If you've tracked Gemini updates closely, like we have, you'll know it's not uncommon for some of the features in the paid tier to trickle down to non-paying users. This typically takes several weeks or even months, in some cases. However, the company is trying a different tactic with Gemini 2.5 Pro Experimental. Related 3 Gemini Advanced features that have made their way to the free tier Advanced features tend to trickle down Posts 3 In a post on X/Twitter over the weekend, the Google Gemini team announced the rollout of Gemini 2.5 Pro (experimental) to all users. The new model is only available on the web version of Gemini right now, though smartphone users can take the new model for a spin by visiting gemini.google.com from the mobile browser of their choice. This is a bit of a surprise, especially given Tuesday's announcement that the new model is only available to users paying for the Gemini Advanced subscription. In the meantime, Google said it is "working hard" to bring the 2.5 Pro experimental model to all smartphone users, though a timeline wasn't provided. Gemini 2.5 Pro is accessible for free, but there are limits in place The reasoning behind Google's change of heart is unclear, though it's quite evident why some Gemini Advanced subscribers would be irked by this. In response to one such user, the Gemini account specified that subscribers also benefit from a larger context window, adding that free users have rate limits on their queries. Gemini's rate limits for each tier and model can be viewed here. This won't be the first time Google has made a paid feature available to non-paying customers, and certainly won't be the last. In February, we saw the file upload feature arrive for users on the free tier following months of exclusivity to Gemini Advanced subscribers. There are many reasons to try out the new 2.5 Pro experimental model, which features better advanced reasoning and coding performance. This model has also exhibited better performance than its industry rivals in common coding, math, and science benchmarks, while debuting at #1 on the LMArena leaderboard.

Google surprisingly rolling out Gemini 2.5 Pro (experimental) to free users

In an unexpected move, Google on Saturday announced that it's rolling out its latest 2.5 Pro (experimental) model to all (free) Gemini app users. Gemini 2.5 Pro (experimental) was just announced on Tuesday as Google's most intelligent AI model that integrates thinking capabilities, with all of the company's future models getting reasoning capabilities going forward. At launch, it was only available for Gemini Advanced subscribers, with Google One AI Premium coming in at $19.99 per month. On Saturday afternoon, Google announced that it has "decided to roll out Gemini 2.5 Pro (experimental) to all Gemini users, beginning today." It's already seeing wide availability on gemini.google.com, and coming soon to mobile. Google says it wants "to get our most intelligent model into more people's hands asap." Gemini 2.5 Pro (experimental) is currently at the top of the LMArena leaderboard, while leading on math and science benchmarks. Google is also working on improving advanced coding performance.

Google's latest Gemini 2.5 Pro AI model is now free for all users

Last week, Google launched the new Gemini 2.5 AI model, the company's "most intelligent AI model" that's designed to tackle increasingly complex problems. At launch, Gemini 2.5 was only available to paying Gemini Advanced and Google AI Studio customers, but that's no longer the case. Over the weekend, Google announced on social media that Gemini 2.5 is "taking off" and the company wants to get the thinking AI model "into more people's hands" ASAP. As such, the experimental version of Gemini 2.5 Pro is now available to all Gemini users, free of charge. However, it remains to be seen how long the AI model will stay free. If you want to try out Gemini's new abilities with Gemini 2.5 Pro, all you have to do is go to the Gemini web app and select 2.5 Pro (experimental) as the AI model from the drop-down menu at the top-left corner of the chat window.

Gemini 2.5 Pro is now free to all users in surprise move

Surprise move comes just days after Gemini 2.5 Pro Experimental arrived for Advanced subscribers. Earlier this week, Google unveiled Gemini 2.5, the company's most advanced AI model to date, with integrated thinking capabilities. It is capable of analyzing complex information with contextual nuance to draw logical conclusions with more accuracy than ever. As users have come to expect, this new model was initially exclusive for those willing to pay $20 per month for Gemini Advanced membership. But less than a week later, Google has surprised everyone by making it immediately available to free users too. "The team is sprinting, TPUs are running hot, and we want to get our most intelligent model into more people's hands asap," the official Gemini App X account wrote in a thread justifying the move. "Which is why we decided to roll out Gemini 2.5 Pro (experimental) to all Gemini users, beginning today." With plenty of Advanced subscribers responding to the post asking what exactly they're paying for if free users get access to the new model too, the official account stressed that Advanced users still get the better experience. Not only do they get a "longer context window", but "free users have rate limits on this model, which do not apply to Advanced users". The official documentation does indeed make it clear that power users will want a paid account. For the Gemini 2.5 Pro Experimental model, Advanced subscribers get four times the requests per minute (20 vs 5) and Requests per day (100 vs 25), as well as double the number of tokens per minute (2 million instead of 1 million). While Gemini's 2.5 model hasn't even been available for a week yet, it is already impressing. It's at the top of the LMArena leaderboard, leapfrogging ChatGPT 4o, and beating its rivals on math, creative writing and science benchmarks. It does, however, lag behind its Open AI rival in terms of coding and following multi-turn exchanges. Impressively, Gemini 2.5 achieved a score of 18.8% on Humanity's Last Exam -- a test designed to push AI's use of complex knowledge to the limit -- putting it ahead of Open AI o3-mini (14%), GPT-4.5 (6.4%), Claude 3.7 Sonnet (8.9%) and DeepSeek R1 (8.6%). That's still a little behind the 26% achieved by OpenAI's Deep Research model, but the comparison is somewhat unfair, as unlike its rivals, it has the ability to search the web, which is useful for the test's general knowledge-based questions.

Gemini's 'most intelligent AI model' yet is now available for free - here are 3 ways you can use its incredible reasoning capabilities

Now you can access its reasoning model, Gemini 2.5 Pro Experimental, for free Google's AI strategy continues to impress, as the company announced this weekend that its brand new Gemini 2.5 Pro Experimental would be rolling out to everyone, not just paid subscribers. "The team is sprinting, TPUs are running hot, and we want to get our most intelligent model into more people's hands asap." The company announced on X, confirming its brand-new reasoning model would be made available for free users, just days after its initial launch. Free users will experience some rate limits compared to paid users, and Google has confirmed that Gemini Advanced subscribers will get 'expanded access and a significantly larger context window.' Gemini 2.5 Pro Experimental is a new AI model with amazing reasoning capabilities that sees it sit at the top of Humanity's Last Exam, one of the hardest AI benchmarks in the world, with 18.8% accuracy. When the new AI model was launched on Wednesday last week, it was only available to Gemini Advanced users who pay a monthly subscription to get the best AI tools Google offers. Google's approach to AI has been incredibly refreshing over the last few months, releasing some of its most powerful tools like Deep Research, an AI agent capable of in-depth analysis, for free. The company has also made a concrete effort to update and upgrade its already available tools continuously, and that's a huge win for consumers. Now that Gemini 2.5 Pro Experimental is available for free, here are 3 ways you can get the most from Google's free new reasoning model. Considering Gemini 2.5 Pro Experimental's impressive result on Humanity's Last Exam, the best place to start with the new reasoning model is by asking it something confusing. Why not try a riddle, or a difficult maths problem to see just how capable Gemini 2.5 Pro Experimental really is? One of my favorite ways to test new reasoning models is by asking it a question without a correct answer. It'll most likely struggle, but it's very interesting to see the thought process to determine just how good the reasoning capabilities actually are. My colleague and AI expert, Eric Hal Schwartz, pitted Gemini 2.5 Pro against ChatGPT o3-mini to find out which AI reasoning model is best. In his experiment, he asked Gemini 2.5 Pro Experimental to develop a web-based application for visualizing the effectiveness of dad jokes, and the results are pretty impressive. While you don't need to recreate a dad joke visualizer, you can use the same technique and prompt to see Gemini 2.5 Pro Experimental's true capabilities, making alterations depending on what you're trying to achieve. Try: "Develop a web application that visualizes the 'success rate' of dad jokes based on various factors. The interface should let users input joke parameters and see projected audience reactions across different demographics. Include elements with playful animations and the ability to save and share your most successful (or painfully unsuccessful) joke formulas." Think of a project that requires lots of logic, something like building a garden extension or starting a sandwich shop business. Gemini 2.5 Pro Experimental excels when it comes to using logic to break down tasks into step-by-step guides, offering detailed instructions on how to achieve your goal. I asked Gemini 2.5 Pro Experimental to create a step-by-step guide for building a cake business, and it created a complex plan with all the information I needed to get my brain ticking. No, Gemini 2.5 Pro Experimental isn't going to start a business for you, but it's clever enough to give you enough basic information to help the logistical side. Gemini 2.5 Pro Experimental won't be for everyone, but the great thing about Google's AI is that you can select the model you need for the task at hand. Now, Gemini 2.5 Pro Experimental is available to everyone, you've got an extra option should you require it.

Gemini 2.5 Pro is now available without limits and for cheaper than Claude, GPT-4o

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Google's Gemini 2.5 Pro, which the company calls its most intelligent model ever, quietly took the developer world by storm. After seeing strong developer interest, Google announced it would increase rate limits for Gemini 2.5 Pro and offer the model at a lower price than many of its competitors. The company did not release pricing at launch. "We've seen incredible developer enthusiasm and early adoption of Gemini 2.5 Pro, and we've been listening to your feedback," Google said in a blog post today. "To make this powerful model available to more developers, we're moving Gemini 2.5 Pro into public preview in the Gemini API in Google AI Studio today, with Vertex AI rolling out shortly." Gemini 2.5 Pro is the first experimental Google model to feature higher rate limits and billing. Google said developers using Gemini 2.5 Pro on public preview, priced at $1.24 per one million tokens, will see increased rate limits. The experimental version of the model will remain free but have lower rate limits. As previously mentioned, Gemini 2.5 Pro is $1.25 per million input tokens and $10 per million output tokens. Social media users expressed surprise that Google could pull off pricing such a powerful model for so low a price, noting that it's "about to get wild." Anthropic offers Claude 3.7 Sonnet, a comparable model to Gemini 2.5 Pro, at $3 per million input tokens and $15 for output tokens. On its site, Anthropic says that Claude 3.7 Sonnet users can save up to 90% of the cost if they use prompt caching. OpenAI's o1 reasoning model costs $15 per million input rockets and $60 per million output tokens. However, cached inputs cost $7.50. Its other reasoning model, o3-mini, is cheaper at $1.10 per million input tokens and $4.40 per million output tokens, but o3-mini is a smaller reasoning model. For non-reasoning models, OpenAI priced GPT-4o at $2.50 for inputs and $10 for outputs. Gemini 2.5 Pro demand Google released Gemini 2.5 Pro somewhat quietly, adding the experimental version of the model to the Gemini Advanced. Since its launch a few weeks ago, several developers and users have found it compelling. VentureBeat's Ben Dickson played with Gemini 2.5 Pro and declared it may be the "most useful reasoning model yet." Effectively pricing reasoning models is the next big battleground for AI model developers. DeepSeek's low cost for DeepSeek R1 caused a ruckus among enterprises. DeepSeek continues to put out models at a lower price than most of the more prominent model developers, putting even more pressure on Google, OpenAI and Anthropic to offer robust and extremely capable models at affordable prices.

Start building with Gemini 2.5 Pro.

We've seen incredible developer enthusiasm and early adoption of Gemini 2.5 Pro, and we've been listening to your feedback. To make this powerful model available to more developers, we're moving Gemini 2.5 Pro into public preview in the Gemini API in Google AI Studio today, with Vertex AI rolling out shortly. Priced competitively, this gives developers access to increased rate limits to use with 2.5 Pro. The experimental version of Gemini 2.5 Pro remains available for free with lower rate limits. We look forward to seeing what you build with 2.5 Pro.

Gemini 2.5 Pro Is Google's Most Powerful AI Model Yet, and It's Already Free

Google has pushed out a shiny new AI model in the form of Gemini 2.5 Pro, albeit with an experimental label next to it -- and it's available for free, so you don't need to subscribe to Gemini Advanced to get it. As with many recent AI model releases, the "reasoning" capabilities of the model are said to be the biggest upgrade here. In artificial intelligence terms, reasoning means answers that are more thoroughly worked through. That should produce fewer mistakes, more logical responses, and a better appreciation of "context and nuance" according to Google. This capability for extra "thought" will now come as standard in future Google models. The Pro (Experimental) release is the first variant of Gemini 2.5 to show up, and while the original blog post didn't mention free users, less than a week later we've got an update saying it's available for everyone -- with rate limits applied if you're not a Gemini Advanced subscriber (Google hasn't specified what those rate limits are). The new model is available now through the desktop app, and coming soon to mobile. Google points to several benchmark tests that show the prowess of Gemini 2.5 Pro. At the time of writing it tops the LMArena leaderboard, where users give ratings on responses from dozens of AI chatbots. It also scores 18.8 percent on the Humanity's Last Exam test -- which measures human knowledge and reasoning -- narrowly edging out rival models from OpenAI and Anthropic. Also of note: the large context window. In simple terms, this is an indicator of how much data the AI model can churn through in one go, and Gemini 2.5 Pro has a context window of one million tokens, with two million "coming soon" according to Google. That compares to a context window of, for example, 200,000 tokens for ChatGPT's o3-mini reasoning model. As tends to be the norm with these AI announcements, there's no mention of copyright infringement as far as training data goes, or increasing energy use. According to MIT researchers, modern-day AI models use a "staggering" amount of electricity and water, and have put us on an "unsustainable path" that needs to change direction quickly. It can be tricky to quantify improvements from one AI model to the next, which is why benchmarks like LMArena are useful. I lack the expert scientific or programming knowledge needed to really put Gemini 2.5 Pro to the test -- though as with the previous model, I was able to create some simple web apps (like an online timer) in minutes. I do know a bit about Charles Dickens' Bleak House, so I set Gemini 2.5 Pro to work on the text. It gave me an accurate summary of the plot, and a clever assessment of the different narrative devices used (which would've really helped me in my study days). It also converted the book into a reasonably well done three-act structure for a movie -- evidence of it holding a lot in its "mind" at once. The older Gemini 2.0 Flash was able to answer the same Bleak House prompts accurately too, but the responses from Gemini 2.5 Pro were longer, more detailed, less generic, and smarter -- evidence of that extra "reasoning" being put to work. The Gemini 2.0 Flash model also had to split the movie adaptation into three responses, perhaps due to the sheer amount of text it was trying to process. Google has provided its own example of the capabilities of Gemini 2.5 Pro, showing how a simple endless runner game can be produced from a single prompt. While the demo video showing the code output is sped up, the game does appear to work and be pretty well-designed, which is an impressive end result from a single natural language prompt. There's also a neat web demo of digital fish swimming around. Elsewhere on the web, the new AI model is being extensively tested. Software engineer and independent AI researcher Simon Willison ran several tests covering image creation, audio transcription, and code generation, and came away very much liking what Gemini 2.5 Pro had been able to come up with. The frenetic pace of AI development shows no signs of slowing down anytime soon, and we can expect more Gemini 2.5 models to appear in the near future. "As always, we welcome feedback so we can continue to improve Gemini's impressive new abilities at a rapid pace, all with the goal of making our AI more helpful," says Koray Kavukcuoglu, from Google's DeepMind AI lab.

Google's Gemini 2.5 Pro could be the most important AI model so far this year

Gemini 2.5 Pro is a "reasoning" model, meaning its answers derive from a mix of training data and real-time reasoning performed in response to the user prompt or question. Like other newer models, Gemini 2.5 Pro can consult the web, but it also contains a fairly recent snapshot of the world's knowledge: Its training data cuts off at the end of January 2025. Last year, in order to boost model performance, AI researchers began shifting toward teaching models to "reason" when they're live and responding to user prompts. This approach requires models to process and retain increasingly more data to arrive at accurate answers. (Gemini 2.5 Pro, for example, can handle up to a million tokens.) However, models often struggle with information overload, making it difficult to extract meaningful insights from all that context. Google appears to have made progress on this front. The YouTube channel AI Explained points out that Gemini 2.5 fared very well on a new benchmark test called Fiction.liveBench that's designed to test a model's ability to remember and comprehend context information. For instance, Fiction.liveBench might ask the model to read a novelette and answer questions that require a deep understanding of the story and characters. Some of the top models, including those from OpenAI and Anthropic, score well when the amount of stored data (the context window) is relatively small. But as the context window increases to 32K, then 60K, then 120K -- about the size of a novelette -- Gemini 2.5 Pro stands out for its superior comprehension.

Google makes its reasoning-optimized Gemini 2.5 Pro model available in public preview - SiliconANGLE

Google makes its reasoning-optimized Gemini 2.5 Pro model available in public preview Google LLC today made Gemini 2.5 Pro, an advanced large language model it debuted last month, available in public preview. Until now, the LLM was accessible through a free application programming interface with low usage limits. Developers could only send 25 requests per day at a rate of up to five a minute. The new public preview version of Gemini 2.5 Pro brings paid tiers that offer significantly higher rate limits. According to Google, developers can now send the model up to 2,000 requests per minute with no daily maximum. Gemini 2.5 Pro is capable of processing up to eight million tokens' worth of data per minute. That makes the model suitable for powering production applications with a large number of users. For prompts with up to 200,000 tokens, Google is charging $1.25 per one million input tokens and $15 per one million output tokens. Those rates increase to $2.5 and $15, respectively, when a prompt's token count exceeds 200,000. Google's pricing makes Gemini 2.5 Pro more expensive than DeepSeek-R1 but cheaper than Anthropic PBC's Claude 3.7 Sonnet. "The experimental version of Gemini 2.5 Pro remains available for free with lower rate limits," Google senior product manager Logan Kilpatrick wrote in a blog post today. At the time of its initial debut last month, Gemini 2.5 Pro topped the popular LMArena LLM benchmark by a significant margin. The benchmark compares AI models' performance based on feedback from users. Gemini 2.5 Pro also achieved a 86.7% score on AIME 2025, a qualifying exam for the U.S. Math Olympiad. Notably, Google says the LLM outperformed several reasoning-optimized models without using "test-time techniques that increase cost." Test-time compute is a machine learning method that boosts an LLM's output quality by increasing the amount of time and hardware resources it invests in completing tasks. The technique can significantly increase inference costs. Under the hood, Gemini 2.5 Pro is improved version of a model called Gemini 2.0 Pro that Google debuted in December. According to the search giant, its engineers enhanced both the base model and the post-training workflow. Post-training is the practice of improving an LLM's output quality after it's trained by providing it with additional data. Google's Gemini model series and open-source LLMs such as R1 are creating more competition for OpenAI. To maintain its market position, the ChatGPT developer will launch a pair of new reasoning models in the next two weeks. OpenAI Chief Executive Officer Sam Altman detailed the plan today in a post on X. The company intends to release the o3 reasoning model it previewed in December, as well as a previously-unannounced LLM called o4-mini. OpenAI originally had no plans to make o3 available as a standalone service. In a few months, the company will follow-up the two models by releasing GPT-5. It's described as an AI system that combines the reasoning-optimized o3 model with several other features. OpenAI will use GPT-5 to power both the free and paid editions of ChatGPT.

Google rolls out experimental version of Gemini 2.5 Pro for free users

Google has launched the experimental Gemini 2.5 Pro model to all free Gemini app users. It features advanced reasoning capabilities, improved coding, and a 1 million token context window. Initially available to Gemini Advanced subscribers and Google AI Studio, it will soon be accessible on Vertex AI.In a surprising move, tech giant Google announced on Saturday that its new Gemini 2.5 Pro (experimental) model will be available to all free users of the Gemini app. The tech giant shared the news on X, stating: "...we want to get our most intelligent model into more people's hands asap. Which is why we decided to roll out Gemini 2.5 Pro (experimental) to all Gemini users, beginning today." However, it's still unclear whether the stable version of Gemini 2.5 Pro will be offered for free to all Gemini users. Initially, it was only accessible to the company's developer platform, Google AI Studio and Gemini Advanced subscribers on the $19.99 per month AI plan. Google Gemini 2.5 Pro: Everything you need to know According to an official blog post, the Gemini 2.5 models are "thinking models, capable of reasoning through their thoughts before responding, resulting in enhanced performance and improved accuracy." Google claims that Gemini 2.5 Pro tops the LMArena leaderboard by a significant margin. LMArena is an open-source platform for crowdsourced AI benchmarking, designed to measure human preferences. The company explained that in the realm of AI, "reasoning" goes beyond just sorting data or making predictions. It's about the system's ability to analyse information, think logically, consider context and details, and make smart decisions. The company is therefore focused on integrating reasoning capabilities into all its models to tackle complex problems and support more advanced, context-aware agents. Gemini 2.5 Pro: Key highlights The model scored 18.8% on Humanity's Last Exam, a test designed to assess knowledge and reasoning skills. The company pointed out its improved coding capabilities, describing it as a "big step up from 2.0," with more enhancements on the horizon. Gemini 2.5 Pro comes with a 1 million token context window. According to news site TechCrunch, this allows the model to process roughly 750,000 words in one go -- that's "longer than the entire Lord of The Rings book series". For developers and enterprises, Gemini 2.5 Pro is available to try in Google AI Studio, with Gemini Advanced users able to access it from the model dropdown on both desktop and mobile. It will also be accessible on Vertex AI soon.

Google releases Gemini 2.5 to the public

While OpenAI servers buckled under massive demand for its new image animation tool, Google launched its most advanced AI brain yet to the public. If you felt that artificial intelligence has been advancing at a breathtaking pace, Google has added more fuel to the fire with the release of Gemini 2.5 to the general public on Sunday. Gemini 2.5, the improved version of its "digital brain," a smarter AI tool with upgraded capabilities, according to the company's claims, is now available for all to experience directly. The announcement came at the end of a weekend when its major competitor, OpenAI, struggled with heavy loads on its servers following a new feature that excited users worldwide. So what does Google promise with Gemini 2.5? According to the company, this isn't just another minor update. The new model is supposed to better understand contexts in long conversations, similar to how a person follows the course of an ongoing discussion. It's expected to be better at analyzing information, writing programming code, and even creative writing, such as poems or short stories. Google emphasized improvements in understanding the subtle nuances of language and reducing the model's tendency to "make up" answers that aren't based on facts -- a well-known problem in the AI world called "hallucinations." The goal, they say, is to make conversations with Gemini feel more natural, flowing, and reliable than before. As mentioned, Google's release came at an interesting time. Just this past weekend, many users who tried OpenAI's new image and animation creation tool encountered difficulties. Thousands in Israel and around the world were excited by the ability to create animated images or complex visuals, such as celebrities in imaginary scenes, and rushed to use the servers. Advertisement The load was so great that it caused significant slowdowns in the company's services, and for brief periods, the service wasn't available at all. The incident illustrated how eager the public is for innovations in artificial intelligence and how intense the competition between tech giants in this field has become. Stay updated with the latest news! Subscribe to The Jerusalem Post Newsletter Subscribe Now Among other things, images could be found on social networks showing leaders who had passed away, such as Begin, Golda Meir, and Ben-Gurion walking on Kaplan Street wearing shirts supporting the release of hostages, images of ultra-Orthodox Jews protesting against military draft dodging while wearing IDF uniforms, and many anime-style animations of everyday life and celebrity scenes. At one point, the company's CEO, Sam Altman, tweeted about how the heavy usage was putting a significant strain on the company's resources, adding, "We need to sleep, too." Sign up for the Business & Innovation Newsletter >>