continue

“Women continue to look for answers outside the health system”

by

For decades, the menopause It has been a topic relegated to the private sphere and, too often, silenced in medical consultations. Given the lack of accessible clinical information, the Internet has become the great refuge for many women who need to resolve their concerns and do not know where to turn. However, what search engines return has changed drastically: health has given way to marketing. What has been seen. A new study published in JAMA has put figures on this phenomenon by analyzing search patterns in Google Trends over two decades in the United States, the United Kingdom and Australia. Based on all this information, researchers have detected an increase of between 15 and 20 percentage points in searches aimed at commercial products and services. That is why the internet is no longer just a place to search for the meaning of a symptom; It is an immense showcase of payment solutions. The health system. This commercial shift in search engines does not occur in a vacuum, but, as Nuria Marín, a menopause specialist, points out, in his analysis for the Science Media Center Spain, the massive increase in these searches reflects a worrying reality: there are fundamental needs not covered in the traditional health system. Women turn to the Internet because they do not find the time, specialized management or comprehensive answers to their routine consultations. However, the study of JAMAdespite being published in a high-impact journal, has a methodological limitation that Marín highlights: Google Trends is an algorithmic and “blind” tool. It does not tell us the age, biological sex, or exact menopausal status of the person typing, which prevents us from establishing direct cause-and-effect relationships at a clinical level. The misinformation. The real problem with seeking answers to questions about our own health is that the algorithm rewards economic profitability over scientific rigor. This means that when a patient searches for information about menopause, she enters a digital ecosystem full of interests. Here is a 2025 study published in B.M.J. revealed that 77.2% of online content on hormone replacement therapy to ‘treat’ menopausal symptoms presented commercial conflicts of interest. But even worse is that 67.2% of the medical claims on these pages They openly contradicted official clinical guidelines based on medical evidence. The quality of the information is quite doubtful, since it was seen that 35% of the websites on menopause had some type of medical quality certification, and more than half required a level of reading comprehension much higher than that recommended for health dissemination. Side B. Not all of the technological landscape is negative, but we have tools that, when designed based on scientific evidence and not aggressive marketing, prove to be a great support for women. Here we have, for example, mobile apps to monitor symptoms which point out that women can see their physical symptoms reduced due to the very fact that the patient structures what is happening to her. That is why the challenge we have ahead of us right now is to offer access to this type of tools and to destigmatize menopause in society. Images | Pexels In Xataka | You get up, you get a cup of hot water, you drink it: more and more people are embracing Chinamaxxing

“Temperatures between 5 or 10 degrees above normal will continue,” says Rodríguez Marcos of AEMET.

by

We’ve known this for weeks. AEMET believes that the June-July-August quarter of 2026 will be in the warm tertile throughout Spain and the truth is that it is not a very risky bet. Of the 10 heat waves that AEMET has been counted in June in the Peninsula and the Balearic Islands since 1975half are concentrated in the last decade. The underlying problem is that, as we are seeing, June has ceased to be a ‘transition’ month to join (statistically and climatically) the summer ‘in a chemically pure state’. Like ‘June’ is ‘summer’? Climatologically it already was, of course. The climatic summer began on June 1. The difference is what the AEMET tells us: the same definition of heat waves was calibrated taking into account the highest temperatures of July and August. That it is increasingly common for June to begin to exceed these temperatures is a clear indication of what we are talking about. Above all, because the wave of June 2025 (from June 18 to July 4) It lasted 17 days and affected 40 provinces: the third longest and the third most extensive of the series. It’s not that this isn’t “the heat from before”, it’s that the situation is out of control. What does this translate into? As explained Francisco Javier Rodríguez Marcos of the AEMET, what the models say is that the temperatures this quarter will be among the 30 highest on record. This will especially affect the northern area, the east of the Peninsula and the Balearic Islands. The line to overcome is clear: the 46 ºC that El Granado (Huelva) reached on June 29 of last year. It is, so far, the highest temperature ever recorded in Spain in the month of June. But in 2026 isn’t that ‘weird’? It is true that we are coming from the wettest winter in 47 years, but it is also true that April was the warmest April in the historical series. That is, no one knows what is ‘normal’ in 2026 and, in this sense, no one can say that the scenario that is drawn is strange. The important thing is to prepare. In some ways, early waves have a potentially greater health impact because the body is not yet acclimated to the heat at the beginning of the season. In this sense, it is best to advance the preparation of the homes (insulation, cooling, night ventilation) and start preparing our daily lives. Image | Tropical TidBits In Xataka | There are areas of Pakistan and India touching 50º C: human beings are discovering where their physiological limit is

Amazon wanted its employees to continue using AI. They have just cut their losses by asking that “you do not use AI just for the sake of using it”

by

Amazon wanted to force its employees They will use AI as if there were no tomorrow. It implemented a tool that measured that usage, but after a few weeks the company realized something: people were using AI for absurd and worthless tasks. That has made Amazon make a decision forceful: abandon this initiative completely. what has happened. Amazon has had to cancel an experiment that measured the performance of its employees based on their use of corporate AI tools. The reason is simple: the engineers had begun to cheat and took the opportunity to automate completely useless and redundant tasks with the sole objective of climbing positions in the ranking. The labor scam has also absurdly increased the computing and infrastructure costs of the company itself, so the experiment has failed. The controversial Kirorank. The service in question was a scoreboard internally named Kirorank. It measured the activity of Amazon developers within Kiro, the “Claude Code of Amazon.” Amazon management wanted 80% of its programmers to use AI every week, an ambitious goal. What the developers ended up doing to score points with their bosses was deploying autonomous agents based on MeshClaw —the version of OpenClaw from Amazon—so that they would run processes in a loop and devour tokens for almost no purpose. The era of tokenmaxxing. Amazon Senior Vice President Dave Treadwell had to intervene this week before the staff to announce that developers no longer had to use this tool. Although he admitted that the experiment had originally been designed with “good intentions,” the practical result ended up being an economic hole due to the tokenmaxxingthat newly coined term that defines the action of artificially inflating the consumption of tokens to simulate productivity. “Please don’t use AI just for the sake of using AI,” the executive demanded of his engineers, urging them to focus on creating better products instead of burning server resources. Cost through the roof. Treadwell’s announcement is no small matter, because this shows that companies have realized that cost control is necessary with AI. Companies like Anthropic—of which Amazon is the largest investor and whose Claude model they use intensively—have recently migrated from flat monthly fees to a per-use pricing model based strictly on token consumption. With this new billing scheme, the fact that the engineers dedicated themselves to “playing” with the bots to rise in the ranking significantly multiplied the bill that Amazon had to pay. Meta suffered the same problem. The Amazon case is not an isolated event. In the Meta and Microsoft offices identical situations have been experiencedwith employees sabotaging internal AI usage rankings through massive token consumption. The irony for Amazon is tremendous: the company has been executing waves of massive layoffs to cut costs and be able to finance its gigantic investment plan in data center infrastructure and AI. Your theoretical capex for 2026 It is estimated at 200,000 million dollars. Lesson learned: AI must be used well. The failure of this “gamification” of work has ended with Amazon abandoning this experiment. To prevent developers from cheating again, a company team is going to change metrics. Instead of measuring raw token consumption, they will analyze so-called “normalized deployments.” From now on, the goal will be to measure how many times the interaction with AI results in useful lines of code that are truly integrated into the company’s products. In Xataka | Customers demand that a human solve their problem. The surprising thing is that if humans serve them they think they are an AI

The million-dollar question in Cupertino is whether Apple can continue being Apple without Tim Cook: Crossover 1×45

by

Tim Cook will stop being CEO of Apple after almost 15 years at the head of the company. It will do so next September 1, the date on which will pass the baton to John Ternusa man of the house with a different career. While Cook has proven to be a genius of logistics and efficiency, Ternus is a man of product and not so much of numbers. This makes us think about the impact that this movement can have from an Apple that in recent years many have criticized for having lost its innovative spirit. The company has shown great success in making the iPhone the absolute center of its strategy, but will that continue to be enough?

While we continue planning how to colonize the Moon, China already has a bricklayer robot to start building a base

by

If we talk about lunar exploration we immediately think of the Artemis programbut the United States is not the only country pushing towards the colonization of our satellite. China also has a program underway and they just showed off a new lunar rover with four wheels and a humanoid upper body. Your job will be to assist in the collection of samples, transportation and deployment of instruments, something like a porter mason. What exactly is it. It is a robot weighing about 100kg with a lower part with four wheels and a humanoid torso with two arms on the upper part. It is not a typical scientific rover, but Its main function is to act as a carrierpicking up and placing different objects and instruments in their positions. The hybrid design, with wheels to move and arms to manipulate, responds to a specific need: on the Moon there are no operators who can move equipment, connect sensors or install instruments. Someone has to do it, and that someone is going to be this robot. Technical challenges. The robot is equipped with AI systems, remote vision and 3D mapping to be able to function in a totally unknown environment. The team that developed it, led by the Hong Kong University of Science and Technology, has South China Morning Post that the main challenge is to ensure that both arms move in a coordinated and precise manner to manipulate fragile instruments. On Earth it is already a complex task, but here you will have to do it in a hostile environment with extreme temperatures, uneven terrain and no one who can repair a possible breakdown. To operate, the robot is powered by solar energy and is designed to operate for two years on the lunar surface, which implies that it will spend 24 lunar nights, each of more than 14 Earth days. During these periods, as it does not receive sunlight, the robot will have to enter a hibernation state and wake up at the beginning of a new day. The mission. The robot is part of the Chang’e-8 missionscheduled for 2028-2029. It will be the eighth mission of the series Chang’e, which China has been using since 2007 to progressively explore the Moon: first orbiters, then landers, rovers and sample collection. The goal of the Chang’e-8 mission is to deliver materials and begin preparing the ground for a permanent presence at the lunar south pole. That’s why the robot is not only designed to explore, but also works. Chang’e-8 is a key part of the International Lunar Research Station (ILRS), the joint project of China and Russia to build a base on the Moon using 3D printing techniques. Why the south pole. The choice of location is not accidental. The lunar south pole has great strategic importance for space agencies because It is where reserves of water ice have been discovered in its craters. That ice has the potential to become fuel, oxygen and water for any permanent base. Whoever arrives first, learns to navigate the terrain and installs more instruments will have a huge advantage. That is why both Chang’e-8 and Artemis III go to the same region. Image | Xataka with Gemini In Xataka | China’s most ambitious space project: an advanced hyperspectral satellite to make a “CT” of the Earth

After a catastrophic 2025, Tesla sales continue to decline in China. The solution is an old acquaintance

by

Sales of electric cars have fallen in China. Although the loss is not as high as that of pure combustion vehicles, the decline in the market is producing very bad results for Tesla. And Elon Musk’s company has brought out one of its traditional tools to boost sales again. An obvious fall. Sales of electric cars in China are not reaping the best results although, everything must be said, recent weeks are beginning to give some hope to companies. At the moment, if global sales are not suffering a setback it is because the accelerator has been put into exportswith record numbers and growth of more than 70% compared to last year. But in the domestic market, sales of “new energy” cars (plug-in hybrids and electric) have fallen 21%reaching 2.92 million cars sold compared to 3.66 million last year. In recent weeks, the Hormuz crisis has served to begin to ground the decline of this type of car. Without state aidits sales had fallen but in recent days we have seen how the savings compared to gasoline have turned the situation around, to the point of break record in plug-in penetrations in the market. Damaged. The context so far this year has not been easy for brands that only sell plug-in vehicles. Much less, therefore, to those who only sell electric vehicles, like Tesla. Without state aid at the beginning of the year and a Chinese New Year longer than usual, sales of this technology fell in a market accustomed to growing year after year. This situation rewarded those who have the most diversified business. In January and FebruaryGeely, which has a portfolio where electric, plug-in hybrids and pure combustion cars are intertwined managed to surpass BYD whose leadership seemed untouchable. Tesla has been through a similar situation. So far this year, Its sales from January to April 2026 have fallen by 15%. It is a bad figure considering all the difficulties the company went through last year. This has led it to lose more market share and remain at just over 3%. Interests. Among the sales of its cars in China, The company has a huge dependence on the Model Ywhich represents around 75% of sales so far this year. But in April, where the Model 3 had a year-on-year drop of 66.09%, the sedan barely accounted for 11% of sales. The fastest solution has been through an old tool: loans. The company has an active campaign in China to defer payments for its cars at 0.99% interest in the case of the Model 3 and 0.92% in the Model Y. The idea is simple, aiming to reward the customer in the long term because it is increasingly difficult for them to compete at the starting price. Right now, in Spain it gives loans above 3% which, however, remains relatively low for our country’s market. However, the company has been offering similar loans before and, right now, In Germany a 0% interest offer is available. Other solutions. Very low interest loans are not Tesla’s only move in China. Aware that the Model 3 has little sales at the moment, GigaShanghai’s exports have skyrocketed so far this year. So much so that global sales, internal sales and those outside Chinese borders, they have grown 36% last April. This means that, clearly, Tesla is trying to move the focus of its target audience. The company has encountered the problem that in China the customer has turned to the local product that usually offers more for less money. The solution is to push the European market, which is now receiving the first units of the Basic Tesla Model 3. less margin. The problem for the company is that it can no longer push the price as hard as before. Before the massive embrace of the Chinese car in its local market and new models began to arrive in the European market, Tesla played as it wanted with demand rising and falling prices. Today those days are over and, what is worse for the company, Your profit margins cannot respond as before. As the price has fallen, the margin has narrowed, losing ability to continue moving in the market. This explains why the voices calling for smaller and more affordable versions of their cars are heard louder. A ship that, given what has been seen, Elon Musk’s company has not been able to bring to fruition. Photo | Priscilla Du Preez and Sou Jest In Xataka | Elon Musk called the $25,000 Tesla an “absurd idea.” Now you need it to compete in China

The spike in Google searches after the 2024 eclipse reveals that we continue to ignore science

by

It has been known for a long time that It is not healthy to look directly at a solar eclipse. It is said that Socrates himself I already recommended looking at it reflected in the waterbut never directly. However, human beings have a fairly significant tendency to ignore scientific recommendations. This is possibly the reason why in 2024, after an eclipse in the United States, Google searches for the phrase “my eyes hurt” had a very abrupt peak. The time and place coincide. That peak of searches took place on April 8, 2024 at 3:00 p.m. Eastern Time. Just at that moment a solar eclipse was occurring whose strip of totality crossed from Mexico to Canada, passing through the United States. The states where the most searches were carried out were Vermont, Arkansas, Michigan, Ohio and Indiana, all of them immersed in the path of the eclipse. Eclipse retinopathy. When looking directly into sunlight, the retina can suffer serious damage. The condition that occurs is known as eclipse retinopathy and causes symptoms such as photosensitivity, blurred vision and headache. Vision can be affected for months or even permanently in the most severe cases. For this reason, no matter how much it may seem like the sun is covered, we must look at it with adequate protection. No sunglasses or x-rays. We have all heard at some point that it is safe to look at an eclipse through an x-ray or a photo negative. However, this is a myth that can be very dangerous. Sunglasses are not safe either. Generally, These are prepared to filter approximately 99.9% of solar ultraviolet radiation.. However, in the event of an eclipse, in which we look directly at the sun, this protection is needed, added to a filter of 99.999% of visible sunlight. It is necessary to use special eclipse glasses, always with filters approved by the competent authorities. Be careful with binoculars and telescopes. We should also not look directly through telescopes or binoculars without using filters. These are placed outside the lens and protect our retinas from solar radiation when we look through them. If none of this seems right to us, we can always resort to a pinhole camera, which reflects the image of the eclipse on another surface. Something like what Socrates advised about looking at the reflection in the water. It is important to use approved glasses You shouldn’t even look at a total eclipse. When the eclipse is total, the Sun is completely obscured. At that point, we might feel safe without protection. The problem is that it is not easy to calculate the exact moment in which the eclipse will begin to dissolve and with just a little bit of light, just when the Sun begins to reappear, we can damage our retinas. It is important to use protection from the beginning. It wasn’t eclipse retinopathy. In reality, the symptoms of eclipse retinopathy They usually appear several hours after the event. Interestingly, eye pain is not one of these symptoms. Therefore, what all those people were looking for was due to another reason. When we look at the sun, we usually experience a blink reflex that forces us to look away. However, with a solar eclipse the brightness is dimmed enough for this reflection to disappear. As a consequence, we can comfortably look at the Sun and keep our eyes fixed, without blinking. That’s what can make our eyes hurt or feel a burning sensation. Specifically, that is not dangerous. Still, those Google searches show that many people were worried. Many of them may not have used protection and regretted it. Ready for August. Next August 12 we will have the first of the eclipses that make up the Iberian Trio. Many people have already bought tickets to travel to some of the points in the totality zone. There are even those who have gotten tickets for one of the many festivals that will be celebrated for this reason. Whatever plan we choose, the important thing is protection. Maybe, even if we protect ourselves, there will be a peak in Google searches, but it better be because we don’t blink for a while and not because we have really damaged our retinas. Image | Magnific/NASA | POT In Xataka | The trio of eclipses that await Spain on the horizon: an unprecedented and historic chain between 2026 and 2028

Samsung, SK Hynix and Micron are already working on DDR6 while memory prices continue to skyrocket

by

The market is going through an unprecedented component crisis. In 2020, a perfect storm caused there to be no chips, but it was known that the storm would pass sooner rather than later. The problem is that the storm of current NAND chip crisis It doesn’t look like it’s going to give us a break in the short term. All memory manufacturers have focused on creating chips for artificial intelligence platformsbut in the meantime, they continue developing DDR6 memories. It will take a long time for consumers to be able to taste them. In short. DDR5 memory has not had a normal life cycle in the consumer market. It was expensive and, when prices began to drop, AI hyperscalers arrived to take over everything available. This is how we have reached a situation in which just 32 GB exceeds the 400 euros: all production is focused on hyperscalers and there are no ‘pills’ for consumption. However, that doesn’t mean that DDR5 has done poorly, not at all. Samsung or SK Hynix they are breaking records and, for them, things are working out great. That is why, together with Micron, they are starting to prepare the ground for the new technology: DDR6 memories. Last year JEDEC (the consortium that is responsible for standardizing DDR, LPDDR and NAND microelectronics) already detailed the standard LPDDR6 and it seems that the fringes are being polished for DDR6. Now, as we read in Wccftechthe big three of memory They are warming up for the near future with their new generation. Performance leap. These new memories will not only be faster. It talks about speeds minimum of 8,800 MT/s according to that JEDEC standard, but can reach up to 17,600 MT/s as the technology develops. They would practically double the performance of DDR5 and the trident of the RAM it takes a few months working together with Intel, AMD and Nvidia in the prototype validation processes. But it’s not just about pure speed, but about architecture. About 2×32-bit DDR5 RAM we would pass to a 4×24-bit subchannel architecture. It is something that brings challenges when it comes to managing temperatures and consumption, but also presents a clear advantage: improved parallelism and greater use of bandwidth. Plate change. For players, that is going to be a little problem, since an architecture change usually entails a board change. And, if the rumors and leaks are true, that plate change will be assured. The reason is that there are sources that they point that CAMM2 is going to gain a lot of weight with DDR6, especially in laptops and compact computers, and could gradually displace the traditional DIMM in certain segments. If this doesn’t tell you anything, visualize how RAM is mounted on a current motherboard. These are modules that are mounted perpendicular to the plate, something that has been around for years and that, although it has been functional for a long time, presents friction when you want to reach certain speeds. On the contrary, we have M.2 SSDs that are mounted in parallel, just like the LPDDR memory of laptops. Precisely, this is what these DDR6 tablets would be like, so manufacturers would have to redesign their boards to adapt either from the front or by adding the connectors to the back of the plate. Context. You probably have two questions in mind: how much DDR6 memory will cost and why the rush. We cannot answer the first question, but for the second question we can guess where the shots are going. We have commented that DDR6 RAM will come hand in hand with a notable improvement in bandwidth, and that is something that the artificial intelligence industry is desperately looking for. Until now, there were many powerful GPUs in data centers to train AI models, but in the era of Agentic AIwhat is needed is equipment more similar to a traditional PC. This is why Intel and AMD sand they are moving to mass-produce their professional processors again, and that is where DDR6 memory would make perfect sense thanks to that improvement in bandwidth. For inference, it’s great. It is already being tested with an arrival on the market for 2028 or 2029, but it will be the hyperscalers who monopolize all DDR6 memory production. Only when the voracity of data centers calms down will the modules begin to reach the mass consumer market. The translation is that Micron, SK Hynix and Samsung are already working on it, but to be able to build a desktop PC with DDR6 there are several years ahead. In the meantime, I’d say we can settle for DDR5, but looking at the prices… Image | Luan Gjokaj (edited) In Xataka | The RAM crisis is destroying all of Valve’s plans with its Steam Machine

If you were waiting for Xiaomi to launch cheap cars, its CEO encourages you to continue waiting seated

by

Xiaomi has been in the automobile market for a couple of years (although it is still we are waiting for your arrival in Europe), and in contrast to what the brand offers in other areas such as smartphones, the company wants to position itself rather high in the price table of its cars. Lei Jun, CEO of Xiaomi, confirmed during a live broadcast on April 17 that the brand has no intention of launching electric vehicles below 100,000 yuan (about 12,500 euros) in the coming years. Here, as expected from the figures, he talks about the Chinese market. Communication. Lei Jun made these statements during a live autonomy test in which he drove a new generation SU7 Pro from Beijing to Shanghai (1,265 kilometers) with a single stop to charge. On the way, he took the opportunity to chat with the chat, a calculated communication strategy that has been noticed. Luckily, during the talk, we were able to find very interesting statements from the head of the brand himself and get an idea of ​​his roadmap. No to the cheap car. According to counted Jun during the broadcast, today’s competitive electric cars increasingly depend on intelligent driving systems, and that type of technology has a high cost that does not fit with a sales price below that barrier of 100,000 yuan in China. According to collect the media CarNewsChina, Lei himself recognized that the new generation of the SU7 It accumulates more than 100 improvements compared to the previous model, with an increase in material costs of almost 20,000 yuan, but its selling price only rose by about 4,000 yuan. For Xiaomi, the equation applied to an entry-level car simply does not add up. Where Xiaomi does want to be. The updated SU7 starts at 219,900 yuan (around 27,500 euros), and the brand’s direction points even higher, as the firm is ready to launch its SU7 Ultra which already competes in the high-performance segment, and in the not too distant future models such as the YU7 GT or a premium variant of the SU7 will also appear, according to they count from ChinaEVHome. We will have to see prices when the firm lands in Europe with its SU7, but everything indicates that Xiaomi wants to consolidate itself within the field of the mid/high range of automobiles. The Chinese car is not synonymous with cheap. Xiaomi is not the only one that avoids the price war in the entry segment. He Xiaopeng, president of XPeng, declared during the presentation of MONA M03 that his company also has no plans to go below that 100,000 yuan threshold. Among the reasons it gave were too tight margins, unsustainable investment in smart technology and real risk of a destructive price spiral. What the numbers say. Sales data in China reinforce this reading. And it is that according to figures collected by CarNewsChina, entry-level electric cars, such as the Wuling Hongguang Mini EV or the BYD Seagull (Dolphin Surf here in Spain), registered year-on-year falls of almost 58% in the first months of 2026, partly due to the end of tax exemptions on purchases. The sedan and utility vehicle segment as a whole also fell almost 20% year-on-year in March. The volume is there, but the profitability is not. Promises. All in all, Lei Jun left a door ajar in the long term. Their goal is for Xiaomi to be among the five largest car manufacturers in the world. Reaching that scale would, sooner or later, require greater price coverage. But for this scenario to come true, there still seems to be time. Cover image | Xiaomi In Xataka | Journey to the center of the Chinese motor (part 2): I have seen the future of cars in Beijing and yes, it is electric (and very cool)

US companies continue to pursue larger and larger AI models. Those from China continue to demonstrate that it is not necessary

by

Until now, Alibaba had a great open model for programming. It is based on Qwen3.5-397B-A17B, but the problem is that it was gigantic with its 397 billion parameters and 807 GB of disk (and memory) size. The Chinese company has done something surprising and has announced these days the Qwen3.6-27B modelwhich in its quantized version weighs less than 17 GB. You would think that at that size he would be much worse than his older brother. But you would be wrong. It is proof that it is possible to give for much less. A dense model. Most large weight models open in 2026 use Mixture-of-Experts architecture (MoE): They have many parameters in total, but only activate a fraction of them when we use them. For example, the Qwen3.5-397B-A17B model precisely indicated that in its name: of the 397,000 million parameters, it only activated 17,000 million (hence the A17B) when using it. With Qwen3.6-27B we have what is called a dense model: the 27 billion parameters are activated in each inference. Although it is somewhat less efficient, it has clear practical advantages. For example, there is no need to configure an expert router, and quantization is more predictable and compact. The idea has worked, and the results prove it. The performance of this “small” AI model is even higher than a much larger previous version. Benchmarks don’t lie (too much). In SWE-bench Verifiedthe most popular benchmark for real programming tasks, Qwen3.6-27B achieves 77.2% score compared to 76.2% for the 397B model. In Terminal-Bench 2.0, which measures how well the model executes tasks in the command console, it achieved 59.3% compared to 2.5% for its rival. But in this test it achieves exactly the same score as Claude Opus 4.5, one of the best recent Anthropic models. That an “Open Source” model that can be easily used locally achieves something like this is unusual, but we must be cautious: the benchmarks are from Alibaba itself, and there is currently no independent verification, although who are wearing they seem be really satisfied with the. Even Alibaba is surprised. What is striking about this launch is that the company that launched it is promoting it above its most ambitious model until recently. Let them compare both versions themselves and recognize that the “small” is the most powerful It is significant. It’s like saying from the rooftops that the largest AI models have no competition, when they have just proven that this is not the case and that models like Qwen3.6-27B can be truly remarkable in behavior. 24 GB of VRAM is “enough”. Thanks to its small size, it is possible to use this model on relatively accessible machines. Thus, the 24 GB of video memory of the RTX 3090 makes these graphics cards a perfect alternative to install and use Qwen3.6-27B with excellent performance. Dense models do not do so well on MacBook or Mac mini with unified memory, and although logically not everyone has access to graphics cards with 24 GB of RAM, access to really capable local models continues to improve. The best essences, in small bottles. Alibaba is a steamroller of “small” AI models, and it demonstrated this in early March when launched several that ranged from 0.8B to 9B. Fortunately there are varied alternatives in that segment of “Small Language Models” (SLMs) and here we have reference examples like Gemma 4just released by Google. Microsoft with Phi-4 (which needs an update, like gpt-oss-20b/120b) or Mistral with Devstral 2 They are examples that Western companies are also making moves in this interesting field. But. According to benchmarks, Qwen3.6-27b is comparable in some benchmarks to Claude Opus 4.5, Anthropic’s most advanced model when it was launched in November 2025. That is surprising and confirms that open weight models from Chinese companies are, as Demis Hassabis saidbetween 6 and 12 months behind the most advanced models from Anthropic, OpenAI or Google. But to execute them a significant investment is still necessary, and although local AI models are very interesting in terms of privacy, if today one wants maximum speed and performance it still depends on commercial models in the cloud. In Xataka | Google will invest up to $40 billion in Anthropic because the new normal for AI is investing in your enemy

Log In

Forgot password?

Forgot password?

Enter your account data and we will send you a link to reset your password.

Your password reset link appears to be invalid or expired.

Log in

Privacy Policy

Add to Collection

No Collections

Here you'll find all collections you've created before.