Tech
Open source Xiaomi MiMo-V2.5 and V2.5-Pro are among the most efficient (and affordable) at agentic ‘claw’ tasks
Xiaomi, the Chinese firm best known for its smartphones and electric vehicles, has lately been shipping some incredibly affordable and high-powered open source AI large language models.
The trend continued today with the release of Xiaomi MiMo-V2.5 and Xiaomi MiMo-V2.5-Pro, both available under the permissive, enterprise-friendly MIT License, making them suitable for use in production in commercial applications. Enterprises and individual/independent developers can now download either of the models (and more Xiaomi open source options) directly from Hugging Face, modify them as needed, and run them locally or on virtual private clouds as they see fit.
The most notable attribute of these models besides the open source licensing is that, according to Xiaomi’s published benchmarks, they are among the most efficient available for agentic “claw” tasks, that is, powering systems such as OpenClaw, NanoClaw and Hermes Agent, in which users can communicate with them directly over third-party messaging apps and have the agents go off and complete tasks on the human user’s behalf, such as making and publishing marketing content, running accounts, organizing email and scheduling, etc.
As Xiaomi’s ClawEval benchmark chart shows, both MiMo-V2.5 and the Pro version in particular appear near the top left of the chart, indicating high performance in completing the benchmarked claw tasks while using the fewest amount of tokens — saving the human user money, especially in a world where more and more services such as Microsoft’s GitHub Copilot are moving to usage-based billing (charging the human behind the agents for each token used rather than imposing rate limits like Anthropic or providing an “all-you-can-eat” buffet-style subscription like OpenAI).
In fact, the Pro model leads the open-source field with a 63.8% success rate, consuming only ~70K tokens per trajectory.
This is roughly 40–60% fewer tokens than those required by Anthropic Claude Opus 4.6, Google Gemini 3.1 Pro, and OpenAI GPT-5.4 to achieve comparable results.
By combining a massive 310B-parameter architecture with a highly efficient “active” footprint and a native 1-million-token context window, Xiaomi MiMo is challenging the dominance of closed-source frontier models from Google and OpenAI, especially when it comes to the latest and greatest craze in enterprise AI deployments — agentic tasks and “claws” similar to OpenClaw.
A two-pronged pincer
Xiaomi has released two distinct versions of the model to serve different ends of the development spectrum: MiMo-V2.5 (the “Omni” multimodal specialist) and MiMo-V2.5-Pro (the “Agent” specialist).
While the base model provides native multimodality, the MiMo-V2.5-Pro is specifically engineered for “long-horizon coherence” and complex software engineering.
On the GDPVal-AA (Elo) benchmark, the Pro model achieved a score of 1581, surpassing competitors like Kimi K2.6 and GLM 5.1.
Xiaomi researchers further released data on several high-complexity tasks performed autonomously by V2.5-Pro:
-
SysY Compiler in Rust: The model implemented a complete compiler from scratch—including lexer, parser, and RISC-V assembly backend—in 4.3 hours. Spanning 672 tool calls, the model achieved a perfect 233/233 score on hidden test suites, a task that typically takes a computer science major several weeks.
-
Full-Featured Video Editor: Over 11.5 hours and 1,868 tool calls, the model produced an 8,192-line desktop application featuring multi-track timelines and an export pipeline.
-
Analog EDA Optimization: In a graduate-level engineering task, the model optimized a Flipped-Voltage-Follower (FVF-LDO) regulator in the TSMC 180nm process. By iterating through an ngspice simulation loop, the model improved metrics like line regulation by 22x over its initial attempt.
These experiments highlight a “harness awareness” in V2.5-Pro, where the model actively manages its own memory and shapes its context to sustain coherence over thousands of sequential tool calls.
Over the API, Xiaomi is pricing the models at competitive rates for both domestic (Chinese) and international markets (like the U.S.). For overseas developers, the high-performance MiMo-V2.5-Pro is priced at $1.00 per million input tokens (for a cache miss) and $3.00 for output within context windows up to 256K.
For ultra-long context tasks between 256K and 1M tokens, the cost doubles to $2.00 for input and $6.00 for output, though the architecture’s caching capabilities offer significant relief, reducing input costs to as little as $0.20 to $0.40 per million tokens upon a cache hit.
Domestically, these rates are mirrored in yuan, with the Pro model starting at ¥7.00 per million input tokens for standard context and reaching ¥14.00 for the extended 1M range. Meanwhile, the base model starts at just $0.40 USD for overseas input per million tokens and $2.00 per million output, putting it among the more affordable third of leading LLMs globally (see our chart below):
To lower the barrier for agentic development further, Xiaomi has made cache writing free of charge for a limited time across all models, alongside a total fee waiver for the entire MiMo-V2.5-TTS suite, which includes its specialized voice cloning and design features.
This pricing logic is clearly designed to accelerate the transition from simple chat applications to persistent, long-horizon agents that can operate at a fraction of the cost of legacy frontier models.
Xiaomi has also introduced an overhauled version of its subscription offerings, called the “Token Plan,” now available in four levels:
-
The Lite “Starter Pack” provides 720 million credits for $63.36 USD per year
-
Standard tier offers 2.4 billion credits for $168.96 per year
-
A Pro tier provides 8.4 billion credits for $528.00 per year (designed for enterprise use cases)
-
Max —aimed at high-intensity coding enthusiasts—delivers 19.2 billion credits for $1,056.00 per year
Beyond credit allotments, all plans include preferential API rates, a 20% discount for off-peak calls, and “Day-0” support for popular coding scaffolds like Cursor, Zed, and Claude Code.
However, both through the API and via the Token Plan, accessing the Xiaomi models from China may present barriers or additional compliance and regulatory risks to U.S.-based enterprise customers. As such, the best bet for U.S. enterprises concerned about relying on Chinese tech but wanting to take advantage of the low cost and open source models is likely setting up their own virtual private clouds or local servers, downloading the model weights, and running the models domestically.
MoE architecture but divergent training regimens for V2.5 and V2.5-Pro
At the heart of MiMo-V2.5 is a Sparse Mixture-of-Experts (MoE) architecture. While the model boasts a total of 310 billion parameters, only 15 billion are “active” during any given inference cycle.
Meanwhile, V2.5-Pro is 1.02 trilion-parameter Mixture-of-Experts model with 42 billion active parameters.
In either case, the design functions much like a specialized research hospital: while the facility has hundreds of doctors (parameters), only the specific specialists required for a particular case (query) are called into the room.
This massive increase in parameter volume for the Pro version provides the “neural capacity” required for the deep, multi-step reasoning found in complex software engineering and long-horizon tasks, as though even more specialists are available in an even larger hospital.
According to Xiaomi’s blog post, the regular V2.5 follows a rigorous five-stage evolution:
-
Text Pre-training: Building a massive language backbone on 48 trillion tokens.
-
Projector Warmup: Aligning in-house audio and visual encoders with the language core.
-
Multimodal Pre-training: Scaling across high-quality cross-modal data.
-
Agentic Post-training: Progressively extending the context window from 32K to 1M tokens.
-
RL and MOPD: Utilizing Reinforcement Learning and Multimodal Preference Optimization (MOPD) to sharpen real-world reasoning and perception.
The backbone utilizes a hybrid sliding-window attention architecture, inherited from MiMo-V2-Flash, which optimizes how the model “remembers” long-range information. This technical foundation enables MiMo-V2.5 to see, hear, and reason natively, rather than relying on external “plug-in” tools for visual or auditory processing.
Conversely, the training of MiMo-V2.5-Pro prioritizes “action space” over sensory perception. Instead of sensory alignment, the Pro model’s training focus shifts toward scaling post-training compute.
This process is designed to instill “harness awareness,” where the model is specifically trained to manage its own memory and context within autonomous agent scaffolds like Claude Code or OpenCode.
While the base V2.5 model is trained to reason across modalities, the Pro version is trained to sustain coherence across more than a thousand sequential tool calls.
The standard V2.5 model balances local and global attention to maintain multimodal perception. The Pro model, however, utilizes an increased hybrid attention ratio—evolving from the 5:1 ratio of previous generations to a more aggressive 7:1 ratio.
This allows the Pro model to “skim” the vast majority of its context while applying high-density attention to the specific 15% of data most relevant to its current objective, a critical feature for debugging large repositories or optimizing graduate-level circuits.
Finally, while both models undergo Reinforcement Learning (RL) and Multimodal Preference Optimization (MOPD), the objectives of these stages differ.
For MiMo-V2.5, the RL stage is used to sharpen perception and multimodal reasoning. For MiMo-V2.5-Pro, RL is focused on instruction following within agentic scenarios, ensuring the model adheres to subtle requirements embedded deep within ultra-long contexts and recovers gracefully from errors during autonomous execution.
This results in the Pro model’s “self-correcting” discipline, as seen in its ability to diagnose and fix regressions during the 4.3-hour SysY compiler build.
Full MIT License is perfect for enterprise use cases
In a move that distinguishes it from many “open” models that include restrictive “Acceptable Use” policies, Xiaomi has released MiMo-V2.5 under the MIT License.The MIT License is the gold standard of permissive software licensing. For developers and enterprises, this means:
-
No Authorization Required: Companies can deploy the model commercially without seeking explicit permission from Xiaomi.
-
Continued Training: Developers are free to fine-tune the model on proprietary data and even release those derivative weights.
-
Unrestricted Commercial Use: There are no revenue caps or user-base limits that often plague “community” licenses.
By choosing MIT over a custom “open weights” license, Xiaomi is positioning MiMo as the foundational infrastructure for the next generation of AI agents, effectively inviting the global developer community to treat the model as a public utility.
Xiaomi’s background: from smartphones and EVs to Chinese open source AI darling
Xiaomi’s pivot toward frontier AI agents is the logical culmination of a decade spent building one of the world’s most dense hardware-software flywheels.
Founded in 2010 as a smartphone disruptor, the Beijing-based company has executed a high-stakes transition into a vertically integrated powerhouse defined by its “Human x Car x Home” strategy. This ecosystem now encompasses over 823 million connectable smart devices unified under the HyperOS architecture.
The company’s 2024 entry into the automotive sector with the SU7 and the subsequent high-performance YU7 SUV served as a proof of concept for this integration, positioning Xiaomi as a direct competitor to global luxury marques.
By investing 200 billion yuan ($29B USD) into foundational R&D for chips and operating systems, Xiaomi has moved beyond consumer electronics assembly; it has become an architect of the “action space,” using its massive hardware footprint as the primary testing ground for the agentic intelligence found in the MiMo-V2.5 series.
Ecosystem support
The release has been met with immediate “Day-0” support from the broader AI ecosystem. The MiMo team announced that SGLang and vLLM—two of the most popular high-throughput inference engines—supported the V2.5 series at launch.
This was made possible through hardware partnerships with AWS, AMD, T-HEAD, and Enflame, ensuring the model can run efficiently on everything from cloud-based H100s to domestic Chinese accelerators.
Fuli Luo, the project lead at Xiaomi MiMo and a former key member of the DeepSeek team, underscored the philosophy behind the release on X (formerly Twitter):
“A model’s value isn’t measured by rankings alone — it’s measured by the problems it solves. Let’s build with MiMo now!”
To kickstart this building phase, Luo announced a 100-trillion free token grant for builders and creators. This massive incentive is designed to lower the barrier to entry for developers who want to experiment with the 1M context window without immediate financial risk.
The economic realignment: open source vs. metered proprietary
The launch arrives at a critical juncture for AI economics. The shift toward usage-based billing marks the definitive end of the “all-you-can-eat” buffet era for AI services, a trend underscored by GitHub’s announcement today that its AI coding assistant Github Copilot will transition all plans to metered, token-based credits.
As seat-based predictability gives way to consumption-driven costs, premium agentic workflows—which can consume millions of tokens in a single reasoning session—are becoming increasingly difficult for enterprises to budget.
User sentiment has turned predictably cynical, with developers lamenting that they will “get less, but pay the same price” as subscriptions convert into finite allotments. This pricing evolution significantly enhances the strategic appeal of the MiMo series. By releasing under a permissive MIT License, Xiaomi allows organizations to bypass the escalating “SaaS tax” and reclaim financial predictability through private deployment.
Crucially, Xiaomi has eliminated the “context tax” for its API. The 1-million-token context window is now billed at the standard rate—1 token = 1 credit for V2.5 and 2 credits for the Pro version—with no additional multiplier. This stands in stark contrast to the industry-wide move toward session-based caps, positioning MiMo as a refuge for cost-sensitive, high-volume development.
Analysis for enterprises
The launch of MiMo-V2.5 is more than just a weight drop; it is a declaration of independence for the open-source community.
By matching Claude Sonnet 4.6 in multimodal agentic work and Gemini 3 Pro in video understanding, Xiaomi has proven that the gap between “closed-door” labs and open research is effectively closed.
With the MIT license as a catalyst and a 100T token grant as fuel, the coming months will likely see a surge in specialized, agentic applications built on the MiMo backbone.
Confirming the project’s ambitious trajectory, the team noted they are already training the next generation, focusing on “deeper reasoning” and “richer real-world grounding”. For now, MiMo-V2.5 stands as a testament to the power of sparse architectures and permissive licensing in the race toward functional AGI.
A modder has turned a Game Boy Color into something you can wear on your wrist, and it’s not just borrowing the look. This is an actual, playable retro console slapped onto your wrist.
YouTuber LeggoMyFroggo managed to squeeze a fully functional Game Boy Color into a wristwatch-sized form factor, creating one of the more bizarre yet impressive retro builds in recent memory.
How’d he cram a Game Boy Color into a tiny watch?
In the YouTube video, modder Chris Hackmann called the project “Time Frog Color”. Rather than going for a simpler route of relying on emulation, the build uses original Game Boy Color hardware, including the Sharp SM83 processor, paired with its video memory and support for physical cartridges.
If that last part sounds insane, it absolutely is. The watch can actually run games using tiny cartridges, which Hackmann even demonstrated by playing Pokémon Gold without any issues. He used an RP2040 chip that handled translating the display signal. This allowed the wearable console to function as a watch when powered off.
How was the gameplay experience?
Shrinking a late ’90s handheld console into a 38mm wristwatch does sound like a cool side project, but it comes with its fair share of compromises. The display is just 1.12 inches, and controls are handled by tiny tactile buttons tucked under 3D-printed caps, which doesn’t exactly sound like game-friendly controls. Making the experience even less immersive is the lack of audio and limited battery life.
In other words, it works, but it’s not exactly the best way to replay your childhood favorites. The Time Frog Color just shows how far retro hardware modding has come. It was never meant to replace the actual Game Boy Color or make gaming on a watch a real thing. Though watching enthusiasts finding ways to preserve and repurpose original components is always fun.
To achieve a smart home, you need a voice assistant to run it. A smart home assistant, usually folded into a smart speaker, will let you command your smart home with your voice and run your various routines. It also acts as a center for every gadget you want to add to your home. And you can add almost anything these days, from smart garage control to even voice-commanding your blinds.
But which assistant should you choose? Each of the big players comes with its own pros and cons, but I recommend choosing based on what you already use day-to-day. Your smartphone is the easiest entry point to pick from Apple or Google, or if you want a huge suite of smart speakers to choose from and have a Prime subscription, you may want to consider Amazon.
Take a look around what’s already in your home to see what works with which ecosystem before deciding. The best system for you will be the path of least resistance, whether that’s using your smartphone’s dedicated assistant or sticking with a platform that best integrates with the devices you already have.
Amazon Alexa
Courtesy of Amazon
WIRED: Huge selection of smart speakers and device compatibility.
TIRED: Paywalls, a meh new assistant, and Ring’s problematic policy.
It all began with Alexa, to some extent. It was the first Amazon Echo speaker back in 2012 that kicked off the smart home in an accessible way, letting anyone voice-command smart bulbs and ask for the weather without needing a custom installer or costing a fortune. Today, Amazon still has the widest range of options. The brand has the most smart speakers by a long shot, with 11 main models of smart speakers and displays currently available, plus several older versions of those same devices also available on Amazon’s website or at other retailers. It’s a huge suite with something for everyone, whether you want a screen, something made for kids, or fantastic sound with Alexa built in.
I do really like Amazon’s speakers and how easy the devices are to use, so this is a great entry point if voice control is of utmost importance. It can bring voice control into any room and for anyone in the house, and Alexa can create different profiles for different members of the family and attach information like calendars to those profiles. Amazon also owns Ring, so those smart home security devices work seamlessly with an Echo speaker, but we don’t recommend using Ring’s cameras because of its partnership with Axon, which enables local law enforcement to request footage directly from Ring users. My colleagues also have concerns about its data collection (and there have been other privacy issues over the years).
You’re also going to hit some paywalls. Amazon has an updated version of Alexa rolling out, Alexa+, which will cost $20 a month unless you have Amazon Prime. (Right now it’s out on Early Access, so it’s free, but non-Prime users can only demo it for 30 days before needing to upgrade to Prime to keep the demo.) The monthly fee is more expensive than Prime membership, so if you want it, it’s better to just join Prime. But neither I nor other WIRED staffers have been impressed by this updated, more expensive Alexa, so I hesitate to say it’s worth any investment. You’ll also need separate subscriptions for Ring devices if you choose to use them.
Alexa Smart Home Starter Pack
Still looking for an Alexa? Here are my favorite devices to start with.
T-Mobile has some cool extras to offer with its home internet and mobile phone service — you can cop some epic streaming perks just for being a customer. Streaming services have been raising prices on a regular basis, so it goes without saying that you may be interested in saving some dollars while still being able to dig into your favorite TV shows, movies, music and podcasts.
T-Mobile is offering customers a slew of solid discounts and freebies that will absolutely save you money while you can still enjoy all the entertainment that comes with subscriptions to Netflix, Apple TV, Paramount Plus and more. I put together a guide below with all the details.
I should note that the discounts listed here are available to T-Mobile customers with eligible home internet and mobile phone plans, unless otherwise specified. Check it out for yourself. Scroll on down for our roundup of the best streaming deals available to you.
Netflix
With eligible mobile phone plans, you can get Netflix with ads for free. If you’ve been wanting to catch up on shows like Bridgerton or The Night Agent, or rewatch hits like Stranger Things, Squid Game and Wednesday, this one is for you. To snag this freebie, you’ll need to have two or more lines with the following plans: Go5G, Go5G Next, Go5G Plus, Magenta, Magenta Max, or any Experience More or Experience Beyond plan. Included in this deal are Military, 55 and First Responder plans, as well.
If you’re an existing Netflix subscriber, you’ll be able to take part in this perk with one of the T-Mobile plans mentioned above. Head to the Add-ons section of your account page to sign up for your Netflix offer.
You can also choose to upgrade and stream ad-free Netflix Premium for $20 per month (down $7 from the regular price) through your T-Mobile bill.
Hulu
If you subscribe to any Experience Beyond or Go5G Next mobile plan, T-Mobile will include ad-supported Hulu for free. One Hulu offer is available per T-Mobile account, and this deal only applies to new and returning T-Mobile customers. To redeem, simply follow the instructions listed here, and you’ll be all set.
Please note that there’s no discount if you wish to switch to a different Hulu plan. And if you’re already paying for Hulu but want to use this T-Mobile freebie instead, you’ll need to cancel your current subscription first.
Paramount Plus
Those of you with an All-In home internet plan can access the Paramount Plus ad-supported Essential Plan at no extra cost. You can watch a range of TV shows (including all of Taylor Sheridan’s hit shows, except Yellowstone) and movies from networks like CBS, Nickelodeon and Comedy Central, plus live sports and a small set of live channels. You can activate the streaming subscription directly through your T-Mobile bill.
SiriusXM
Do you have a T-Mobile wireless plan, such as Experience Beyond, Experience More or the Go5G plans? Well, you can get the SiriusXM All Access plan (for app use only) free for six months. To snatch up this deal, simply add it to your account. After the six-month promo is up, customers will be billed the full price for the service, currently $12 per month.
New and returning SiriusXM subscribers who’d like to use this perk can, as long as you haven’t had an active subscription to the satellite radio service in the past 12 months.
Pandora
Speaking of streaming music, four months of Pandora Premium is included for free if you’re a T-Mobile postpaid mobile customer. You can add the Pandora Premium On Us perk to your account through the T-Mobile website or app to take advantage of the deal. After the four months are up, your bill will show Pandora Premium’s regular price, which is currently $11 a month.
Apple TV isn’t free, but it’s cheap
Since 2021, T-Mobile has offered its subscribers complimentary access to Apple TV. That deal ended on Jan. 1. Customers with premium mobile and voice plans saw the free perk replaced by a $3-per-month cost for the streamer. T-Mobile newbies can also sign up for the deal.
This change reflects Apple TV’s recent price bump from $10 to $13 per month. Anyone with the 55+ and Senior plans, as well as phone plans for military members, first responders, and individuals with hearing or visual impairments, will receive this $3 deal for six months. While not free, this new price is still a noteworthy discount for the streamer.
The robots will be introduced at the start of May by Japan Airlines on a trial basis, though the ultimate goal is to deploy them permanently. If you’re one of the 60 million people passing through Haneda airport every year, keep a lookout for one.
Read Entire Article
Source link
Some films age quietly. Others sit in the dark, waiting for the format to catch up. The Man Who Wasn’t There lands in the second category. Shot in black and white by the Coen Brothers, this 2001 noir has always been more about mood and control than plot momentum, and the jump to 4K from the Criterion Collection finally gives its visual look the kind of presentation it needed.
The timing isn’t accidental. Billy Bob Thornton is back in the conversation thanks to his performance in Landman, and it’s a reminder that his turn as Ed Crane remains one of the most restrained and quietly devastating performances of his career. He barely raises his voice, barely moves the needle emotionally on the surface, and still manages to pull the entire film into his orbit. It’s control bordering on suffocation.
Across from him, Frances McDormand does what she always does; make it look easier than it is. She’s sharp, cynical, and completely believable, even if her work here doesn’t quite reach the level of Fargo. And then there’s James Gandolfini, gone far too soon, reminding everyone that while The Sopranos defined him, it never boxed him in as an actor.
This isn’t a noir that grabs you by the collar. It just stands there, lights a cigarette, and lets the smoke do the talking. Compared to sharper, more aggressive classics like The Asphalt Jungle, Double Indemnity, or even Out of the Past, it doesn’t have the same edge or narrative snap. There’s no real jolt, no clever turn that resets the stakes and forces you to lean forward. Instead, it moves at its own pace, more interested in mood and control than tension.
That’s not a flaw, but it does change how it lands, especially if you’re expecting the kind of bite those earlier films delivered. The Coen Brothers aren’t Billy Wilder. Not even a distant relative with a suspicious accent and a better third act.
Ed Crane is a quiet barber who suspects his wife is having an affair. Rather than confront it, he tries to use the situation to fund a small business opportunity through blackmail. It seems simple enough, but the plan quickly leads to complications he doesn’t fully understand or control.
It goes wrong in the way these things always do. Not all at once, and not with much warning. One decision leads to another, each one a little worse than the last, until Crane is in over his head and still acting like he has a handle on it. He doesn’t. And by the time that becomes obvious, it’s already too late.
Image & Sound Quality
Criterion keeps this one simple. The 4K Ultra HD disc is region-free, the included Blu-ray is Region A locked. The new restoration comes from the original 35mm camera negative and is presented in native 4K with Dolby Vision and HDR. I watched most of it in Dolby Vision and then checked in on the Blu-ray to see how much you’re really missing.
Not a lot, but it’s there.
The image looks clean and properly resolved without any heavy-handed processing. Grain is present and stable, detail is consistent, and the overall presentation feels natural. Depth is better than expected for a black-and-white title, with solid separation between foreground and background elements.
Grayscale is the real strength. Blacks are stable, whites stay in check, and the midtones carry the weight without getting muddy. Dolby Vision helps a bit with control, but this isn’t a dramatic HDR showcase. It’s more about refinement than range.
The standard Blu-ray holds up well. On a modest setup, you could live with it and not feel shortchanged. On a larger screen, the 4K disc has the edge; slightly better clarity, a bit more stability, and cleaner fine detail. Most certainly one of their best reissues in awhile in the genre.
Audio is limited to a single English DTS HD Master Audio 5.1 track with optional English SDH subtitles. The original 5.0 mix is presented in a 5.1 format, but this is not a film that makes aggressive use of surround channels. It is primarily dialogue-driven, with a restrained sound design that reflects the Coen Brothers’ usual approach.
Dialogue and narration are clear and easy to follow throughout. The track handles quieter scenes well, where small shifts in volume and tone are more noticeable than any large dynamic moments. There are a few louder sequences that open things up slightly, but they are not the focus.
Criterion splits the extras across both discs, with one key item carried over on each.
Both the 4K and Blu-ray include the same archival commentary featuring Joel and Ethan Coen and Billy Bob Thornton, recorded in 2004. It’s a measured track that focuses on structure, tone, and character behavior rather than production trivia. There’s also some discussion about visual choices and how certain scenes were shaped, which ties back to the film’s overall look.
The Blu-ray adds the rest of the material. The new 37-minute conversation between the Coens and critic Megan Abbott is the most substantial piece. It covers the film’s origins, its place within noir, and how they approached its restrained style. It’s direct and stays on topic.
There’s also an older 13-minute segment with Roger Deakins that focuses on the cinematography. He walks through the visual approach and some of the decisions behind the black-and-white presentation.
The remaining extras are brief. A 10-minute behind-the-scenes piece offers raw footage from the set without much structure. Two short deleted scenes are included but don’t add much context. The package also comes with a printed leaflet featuring an essay by Laura Lippman and standard technical notes.
Overall, the extras are focused but not extensive. The commentary and the new Coen/Abbott discussion carry most of the weight.
Movie Details
- STUDIO: Criterion
- FORMAT: Ultra HD 4K Blu-ray
- THEATRICAL RELEASE YEAR: 2001
- ASPECT RATIO: 1.85:1
- HDR FORMATS: Dolby Vision, HDR10
- AUDIO FORMAT: DTS HD Master Audio 5.1, English SDH Subtitles
- LENGTH: 116 mins.
- MPAA RATING: R
- DIRECTOR: Joel Coen, Ethan Coen
- STARRING: Billy Bob Thornton, Frances McDormand, Michael Badalucco, James Gandolfini, Katherine Borowitz, Jon Polito
Our Ratings
★★★★★★★★★★ Picture
★★★★★★★★★★ Sound
★★★★★★★★★★ Extra
Where to buy:
Related Reading:
Snabbit, an Indian on-demand home services startup, has closed a $56 million funding round, confirming TechCrunch’s earlier report.
Co-led by Susquehanna Venture Capital, Mirae Asset Venture Investments’ Unicorn Growth Fund, and Bertelsmann India Investments, the company’s Series D round values the Bengaluru-based startup at around $350 million, according to a person familiar with the matter. That’s up from $180 million about six months ago. Existing investors Nexus Venture Partners and Lightspeed also participated, alongside FJ Labs. The company has raised about $112 million in total.
Founded in 2024, Snabbit said it is now processing over 40,000 jobs daily across a network of more than 15,000 workers in five cities, offering services such as cleaning, dishwashing, and laundry as demand for rapid, on-demand home services grows in urban India.
The startup said the amount it loses on each order has fallen about 50%, while its customer acquisition costs have shrunk roughly 65%.
Snabbit’s fundraise comes as investor interest in India’s on-demand home services sector heats up, with rival Pronto also in talks to raise fresh capital and publicly traded Urban Company reporting more than one million monthly bookings.
While Apple’s keynote video is typically the most memorable part of WWDC, Apple has shown off how and where developers can participate in a bunch of community events across the globe before and after the conference.
More than 20 community-driven events will be held before and after WWDC 2026.
Apple’s Worldwide Developers Conference is held in June each year, at Apple Park in California. The company uses its annual event to preview new versions of its major operating systems, with iOS 27 and more expected to debut on June 8.
In-person WWDC attendees are selected through a lottery program, but even those who didn’t win will have access to conference-adjacent and community-hosted events.
Continue Reading on AppleInsider | Discuss on our Forums
If that one big sum per year for an annual app subscription is too much to pay at once, the App Store will let you break that into monthly payments.
Apple is making it easier to get annual subscription discounts
Subscription revenue has become an integral part of Apple Services. Developers have a few ways to draw customers into long-term usage and bigger payouts, but they require big up-front financial commitments.
Apple has revealed a new payment option that could help developers offer better discounts while still getting a long-term commitment from the user. It works by offering the user an annual discount that is paid for over 12 months.
Continue Reading on AppleInsider | Discuss on our Forums
Even Realities has rolled out the v2.2.0 update for its Even G2 smart glasses, with Terminal Mode as the headline feature. The update is now live, giving developers a new way to monitor and interact with coding agents without staying glued to a laptop screen.
Your terminal has entered your eyeballs
Terminal Mode is built around the idea of putting the coding terminal in front of your eyes. Developers using AI coding agents can track what the agent is doing, check progress, give commands and respond when needed through the glasses’ built-in microphone. The feature reduces the need to keep jumping back to a laptop, giving users more freedom to move around, handle chores, or even work out between coding sessions.
CEO Will Wang said in an interview that the idea for this feature came after a recent visit to Silicon Valley, where he noticed developers increasingly speaking to AI agents through microphones instead of typing every command. These agents can now complete most of the work on their own, which made Even G2’s built-in microphone and virtual terminal display feel like a natural fit for this workflow.
Sure, it’s nerdy. But so is wearing a computer on your face. Terminal Mode leans into what G2 already does well, surfacing just enough information to keep you moving, and applies it to code.
Terminal Mode fits into Even Realities’ push to make the G2 useful for more than basic notifications. The glasses recently added tools for meeting prep and running apps directly in front of the wearer’s eyes.
The nerdy stuff does not stop there
The v2.2.0 update also includes a few other additions. Conversate and Translate records can now be exported. Even AI now supports continuous conversations, so users no longer need to repeat a wake-up command during an ongoing interaction. Even Realities also mentions sleep algorithm improvements and better connection stability.
Terminal Mode may sound like a feature built for a very specific crowd, but for coders already using AI agents, it could become a useful second screen for quick updates, approvals, and progress checks. It is a niche feature for now, but it gives the Even G2 a clearer role in developer workflows beyond basic smart-glasses notifications.
Although it’s currently only available to buy in Malaysia, we’re keen to see how the top-end Honor 600 Pro compares to the Samsung Galaxy S26.
Ahead of our review, we’ve compared the specs of the two Android phones and highlighted all the noteworthy differences between them below. Keep reading to learn more about what separates the Honor 600 Pro from the Samsung Galaxy S26.
Not sold on either handset? Our best Android phones and best mid-range phones should have you covered.
Price and Availability
At the time of writing, the Honor 600 Pro is only available to buy in Malaysia although it will see a more global launch in the coming weeks.
The Galaxy S26 is part of Samsung’s 2026 flagship S-series. With a starting RRP of £899/$899, the S26 is the most affordable of the line-up.
Snapdragon 8 Elite vs Exynos 2600
Samsung caused something of a stir when the S26 series first launched, as it was revealed that while US customers see Snapdragon 8 Elite Gen 5 for Galaxy across the entire line-up, elsewhere the Galaxy S26 and S26 Plus are fitted with Exynos 2600.
Even so, Samsung promises there shouldn’t be any differences in performance between Exynos 2600 and Snapdragon 8 Elite Gen 5. In fact, we found that the phone benchmarked strongly against the Snapdragon 8 Elite Gen 5-equipped Ultra variant.
Otherwise, and as expected, the S26 runs brilliantly for everything from scrolling and taking photos to even intense gaming sessions too. So, although it may not boast a Qualcomm chip, it’s still a solid performer.
Instead, Honor 600 Pro runs on Qualcomm’s Snapdragon 8 Elite – the brand’s 2025 flagship. With this in mind, while it will likely struggle to reach the heights of Snapdragon 8 Elite Gen 5 and Exynos 2600, it is still a great chip that powered many of the best Android phones last year.
Honor 600 Pro has a larger battery
Samsung isn’t known for equipping its phones with mighty batteries, and the Galaxy S26 is no exception. In fact, with a 4300mAh cell, it’s actually on the smaller side. Having said that, we still found the S26 to be a solid all-day phone, especially for those who average around three to four hours of screen time a day.
In comparison, the Honor 600 Pro sports a mammoth 7000mAh cell. We’ll have to wait until we review the Honor 600 Pro to see how its battery life really measures up, however we’d hope that such a large cell will result in at least a full day, or even a two-day charge.
Honor 600 Pro is IP68, IP69 and IP69K rated
Durability should be a key consideration when you’re buying a new phone, and it’s fair to say that Honor doesn’t want you to take any chances with the 600 Pro. In fact, with IP68, IP69 and IP69K ratings, not only is the handset dust-tight but it can survive water submersion and even exposure to high pressure and high temperature water jets too.
That undoubtedly sounds impressive on paper, but we’d argue that perhaps IP69 and IP69K ratings aren’t that necessary. After all, how often will your phone realistically be exposed to water jets?
Samsung takes a more realistic approach, as the Galaxy S26 is equipped with a simple IP68 rating instead. That means it’s dust-tight and can survive water submersion too.
Honor 600 Pro has a 200MP main lens
Both phones are fitted with three rear lenses, including a main, ultrawide and a telephoto, however they differ with their exact resolutions and offerings.
The Galaxy S26’s camera hardware may seem a bit familiar, as Samsung hasn’t made any major changes in the last few series. This is a shame as, although overall the camera hardware is solid, it’s starting to show its age – especially as it needs to compete with the likes of the best camera phones.
While the main 50MP lens is the strongest of the three, and can cope across most lighting conditions, the ultrawide is easily the weakest of the bunch. It’s fine in bright conditions, but image quality drops quickly in difficult lighting situations, with images looking grainy and rough. Finally, the 3x telephoto works well in specific scenarios, but it can struggle to completely lock focus with an object.
The Honor 600 Pro is instead fitted with an impressive sounding 200MP main camera that’s supported by a 50MP 3.5x zoom and a 12MP ultrawide. As we’re yet to review the Honor 600 Pro, we’ll have to wait and see how the hardware really performs. However, Honor promises that the camera set-up should offer true-to-life colour reproduction, impressive stability and better shots taken at night too.
Honor AI vs Galaxy AI
Is it even a smartphone in 2026 if there isn’t a sprinkling of AI features? Both the Honor 600 Pro and Galaxy S26 are equipped with plenty of AI tools, including access to Google Gemini too.
You’ll likely have heard of Galaxy AI, as the toolkit is arguably one of the most fleshed out available, with genuinely useful features including an array of photo editing capabilities, Live Translate and more.
In comparison, the Honor 600 Pro has its own dedicated AI button that opens up what’s arguably Honor’s headline feature: AI Image to Video 2.0. This tool allows you to turn up to three images into a short video, with just a few prompts.
Otherwise, like Galaxy AI, there are photo editing features to remove unwanted objects from images, plus the Honor 600 Pro includes AI scam and deepfake detection too.
Early Verdict
It’s difficult to compare the Honor 600 Pro and Samsung Galaxy S26 fairly, as we don’t know how much the former will cost in the UK. Having said that, if you want a phone with a larger battery, a fairly recent Qualcomm chip and an intriguing camera set-up, then the Honor 600 Pro could be the one for you.
On the other hand, if you want to play around with a more established set of AI features, want a reliable camera set-up and are in Samsung’s ecosystem, then the Galaxy S26 remains a tough one to beat.
We’ll be sure to update this versus once we review the Honor 600 Pro.
Full-Cycle Software Development: From Idea to Product
Is Whales Accumulating WOJAK at a $30 Million Market Cap: Is Crypto’s Most Iconic Meme Coin About to Explode?
10 Greatest Sci-Fi Thrillers of All Time, Ranked
Manchester United reach agreement with Casemiro over contract clause amid transfer speculation
US brings back mandatory military draft registration
Steven Gerrard disagrees with Gary Neville over ‘shock’ Chelsea and Arsenal claim | Football
-
Fashion4 days agoWeekend Open Thread – Corporette.com
-
Tech18 hours agoRegister Renaming | Hackaday
-
Crypto World3 days agoHyperliquid $HYPE Rally Builds Momentum as AI Sector Enters Prove-It Phase
-
Politics6 days agoMaking troops accountable for war crimes threatens US alliance, ex-SAS colonel warns
-
Politics6 days agoDisabled people challenge government SEND proposals over segregation concerns
-
Business4 days agoPatterson-UTI Energy, Inc. (PTEN) Q1 2026 Earnings Call Transcript
-
Business6 days agoRolls-Royce Voted UK’s Most Iconic Trade Mark as IPO Register Hits 150
-
Crypto World7 days agoFive Value Stocks with Recovery Potential in 2026: PayPal (PYPL), Nike (NKE), and More
-
Sports2 days agoIPL 2026: Ruturaj Gaikwad registers slowest fifty of the season, enters all-time unwanted list | Cricket News
-
Politics14 hours agoDrax board avoid their own AGM, accused of greenwashing & environmental racism
-
Crypto World7 days agoNew York sues Coinbase, Gemini over prediction market offerings
-
Politics6 days agoStarmer handler McSweeney to be dragged from shadows by Foreign Affairs Committee
-
Politics6 days agoZack Polanski responds to home secretary’s taser threat
-
Politics6 days ago
Wings Over Scotland | How To Get Away With Crimes
-
Entertainment7 days ago
Sydney Sweeney cameo cut from “The Devil Wears Prada 2”, source explains why (exclusive)
-
Business6 days agoHCL Tech share price tank over 9% after weak Q4. JPMorgan, HSBC & 3 others cut target price
-
Crypto World7 days agoCrypto’s great hope in Senate’s Clarity Act still has a path to survive tight calendar
-
Politics6 days ago‘Iran is still a nuclear threat’
-
Fashion7 days agoKilkenny Design New Beauty Arrivals for Spring 2026
-
Sports5 days agoTim Bradley names the current best in the world: “Better than Inoue and Usyk”
You must be logged in to post a comment Login