Who has the most advanced humanoid robot?

Why can't we simply rank the best humanoid robots?

We cannot produce a clean humanoid robot league table today because the leading companies still choose their own tests.

NIST recently pointed out that the DARPA Robotics Challenge, held more than a decade ago, was the last serious event where several humanoids were measured side by side. Its proposed new benchmark would test walking, manipulation, coordinated movement and basic reasoning, but that common test is still being developed.

For now, every company shows the part of the problem it handles best. Boston Dynamics publishes heavy lifting, extreme reach and unusual body positions. Figure shows long household sequences and learned factory work. Agility counts totes moved in a live warehouse. Unitree highlights running, recovery and price. Tesla talks increasingly about future production.

The videos also hide different amounts of help. A task can be fully autonomous, teleoperated, replayed from recorded motion, guided by a human at key moments or assembled from several learned skills. Those distinctions often disappear once a short clip spreads online.

We therefore give the most weight to repeated work, stated autonomy conditions, useful task variety and hardware that can survive outside a lab. A one-off trick can reveal technical depth. It cannot settle the ranking.

What does "most advanced humanoid robot" actually mean?

For this question, the most advanced humanoid is the robot that best combines autonomy, body control, useful hands, reliability and repeatable manufacturing.

A humanoid needs enough strength and balance to move through spaces built for people. It also needs to understand changing scenes, handle objects without constant retraining and recover when a grasp or step goes slightly wrong. Hands deserve special attention because most valuable human work ends with touching, holding, turning, pulling or fitting something.

Reliability separates a research success from useful equipment. Ten polished minutes tell us less than hundreds of ordinary hours with jams, shifted objects, warm motors and worn components. Production counts too. A company learns much faster when it can build hundreds of comparable robots instead of nursing a few custom machines.

Some things count more than others here. The biggest test is whether the robot can think and move its whole body through a changing task. Hands and reliability come next. Factory scale only helps once the machine is useful.

What we measure	Strong evidence	Why it changes the ranking
Autonomous intelligence	Long tasks, changing objects, few human interventions	Shows whether the robot can adapt rather than replay motions
Whole-body control	Walking, balancing and manipulating at the same time	Reveals whether the humanoid shape adds real capability
Hands and manipulation	Two-handed work, touch, force control and object variety	Determines how much useful work the robot can actually perform
Reliability	Reported hours, cycles, failures and customer-site operation	Exposes weaknesses that short demonstrations hide
Manufacturing	Hundreds of comparable units, measured yields and fleet support	Shows whether progress can spread beyond a small lab fleet

Which humanoid robots are genuinely in the race today?

Figure, Boston Dynamics and Agility form the top tier, with Tesla, Unitree, UBTECH, 1X and Apptronik chasing from different directions.

Figure looks strongest overall. Helix 02 links vision, reasoning, walking, balance and manipulation, while Figure 02 brought back substantial factory data from BMW. Figure has also produced more than 350 units of Figure 03 and recently returned that model to BMW for a harder sequencing workflow.

Boston Dynamics has the most capable body. The production Atlas combines a 50-kilogram peak lifting capacity, a 2.3-metre reach, continuous rotational joints and industrial environmental protection. Its public customer record is younger than Figure's or Digit's.

Agility's Digit has performed the most clearly measured commercial work. It passed 100,000 tote movements at a GXO facility and has spent far longer inside a live logistics workflow than most rivals.

The chasing group contains serious threats. Unitree sells capable hardware at prices that open the field to universities and developers. UBTECH has moved Walker S2 into production and industrial delivery. 1X has started producing a home robot with unusually advanced hands. Apptronik is building large robot training facilities with Google DeepMind. Tesla has the clearest route to enormous manufacturing scale if Optimus reaches the required capability.

Robot	Best evidence	Main gap	Our position
Figure 03	Broad learned autonomy, BMW experience and a large new fleet	Most new work remains company-reported	Overall leader
Boston Dynamics Atlas	Best hardware specifications and industrial design	Limited published production throughput	Closest challenger
Agility Digit	Sustained commercial throughput at GXO	Narrower manipulation and task range	Worksite leader
Tesla Optimus	AI talent and a planned high-volume line	Little comparable autonomy or customer data	High-upside follower
Unitree H2/G1	Dynamic hardware, open sales and strong value	Weak public evidence of sustained useful work	Hardware ecosystem leader
UBTECH Walker S2	Industrial access, production and factory software	Few comparable runtime totals	Leading Chinese industrial contender
1X NEO	Home design and production underway	Basic autonomy still needs remote help	Home-product leader
Apptronik Apollo 2	Large data operation and Google DeepMind partnership	Commercial results remain lightly quantified	Strongest sleeper

If you want more recent data on this point, please see our latest humanoid robotics market report.

Who has built the best humanoid robot body?

Boston Dynamics has built the best humanoid body we can assess: Atlas is stronger, tougher and more mechanically free than its peers.

The published specifications are unusually convincing. Atlas stands 1.9 metres tall, weighs 90 kilograms, sustains a 30-kilogram load and can briefly handle 50 kilograms. Its 56 degrees of freedom and 2.3-metre reach let it work near the floor, above shoulder height and around awkward containers. Continuous rotational joints allow the robot to turn limbs through positions that would force a more human-like machine to reset its stance.

Atlas is also designed for rougher workplaces. Boston Dynamics lists an IP67 rating and an operating range from minus 20 to 40 degrees Celsius. Its four-hour battery sounds ordinary until the robot walks to a station and replaces the pack by itself. UBTECH's Walker S2 offers a similar three-minute battery swap, but Atlas combines continuous operation with greater strength, reach and joint freedom.

Unitree remains impressive on dynamic movement. Its H2 has 31 degrees of freedom and joint torque reaching 360 newton-metres, while the cheaper G1 has made capable humanoid hardware far easier to buy. Those platforms are excellent research machines. Atlas looks more ready for industrial work.

Boston Dynamics also brings decades of experience with balance, falls, whole-body control and mobile robot servicing. For a physically demanding job where hardware limits matter more than broad reasoning, Atlas is the obvious first choice.

Which humanoid robot has the smartest AI today?

Figure 03 currently has the best public record for the smartest full humanoid because Helix 02 controls walking, balance, vision and manipulation together.

Figure says Helix 02 sends onboard vision, touch and body-position data through one learned system that controls the robot's full set of actuators. Its lower-level controller was trained with more than 1,000 hours of human motion data plus reinforcement learning in simulation. That setup lets the robot adjust its feet, torso, arms and hands as one action unfolds.

The four-minute dishwasher task remains one of the clearest examples. Figure 03 crossed a full-sized kitchen, handled different dishes, kept its balance and completed the sequence without a reset or human intervention. A more recent bedroom demonstration pushed the idea further: two robots opened doors, moved a chair, handled clothing, used a foot pedal and made a bed together in under two minutes. Figure says they shared the same learned policy and inferred each other's movement without a central coordinator.

These are company-selected demonstrations, not neutral benchmarks. Still, the pattern has widened over time. The same core approach has appeared in logistics, laundry, kitchen work, living-room cleanup, bedroom work and now BMW sequencing. That repeated transfer makes a stronger AI case than one spectacular clip.

Boston Dynamics may close the gap quickly through Atlas and its Google DeepMind partnership. Apptronik is feeding Apollo 2 data into Gemini Robotics, while NVIDIA's GR00T platform is giving many manufacturers access to better foundation models. On the evidence available, Figure has shown the most convincing connection between a general learning system and a capable full body.

If you want more recent data on this point, please see our latest humanoid robotics market report.

Who has the best humanoid robot hands?

Figure leads demonstrated humanoid dexterity, although 1X has just raised the hardware ceiling with NEO's new tendon-driven hands.

Figure's advantage comes from what the whole robot has handled. Its systems have worked with boxes, thin sheet metal, dishes, towels, clothing, bedding, remote controls, carts and other objects that require different grips and force levels. At BMW, Figure 02 placed sheet-metal parts within a five-millimetre tolerance and targeted more than 99% success for a full three-part cycle during a shift.

The newer household work adds bimanual and deformable-object control. Figure 03 has reoriented headphones in mid-air, closed a book with two hands, operated a trash-can pedal while balancing on one leg and coordinated with another robot on a comforter. Those jobs combine hand control with vision, foot placement and body posture.

1X now has the most interesting hand design on paper. NEO's new tendon-driven hands have 25 degrees of freedom, backdrivable joints and tactile skin intended to sense pressure and slip. The company says these hands will ship on every NEO. That design could eventually beat Figure on fine touch, safety and human-like range.

The proof is still catching up with the specification. 1X has shown the hardware and described its sensing architecture; Figure has published a broader record of autonomous manipulation. Figure keeps the lead for now, while 1X has the clearest path to taking it.

Which humanoid robot has done the most real work?

Digit remains the most proven paid humanoid worker, while Figure 02 completed the more delicate factory job.

Agility Robotics reported that Digit moved more than 100,000 totes at GXO's Flowery Branch facility. The robot worked inside an ongoing commercial deployment, transferring containers between autonomous mobile robots, conveyors and floor positions. Six figures of repeated movements expose navigation errors, grip failures, battery interruptions and workflow problems that a demonstration can avoid.

Figure's BMW record covered a smaller number of handled items but demanded more precision. Figure 02 accumulated over 1,250 hours and loaded more than 90,000 sheet-metal parts. It also walked an estimated 1.2 million steps and contributed to the production of more than 30,000 BMW X3 vehicles. The robot had to place three parts within a five-millimetre tolerance while meeting an 84-second cycle requirement.

The numbers produce two winners. Digit has the stronger commercial reliability story because its tote workflow crossed a clear throughput milestone and stayed active long enough to become ordinary warehouse work. Figure's deployment says more about dexterous factory integration and learning from hardware failures.

The next competitors remain far behind on published totals. UBTECH has entered several vehicle factories, Atlas has begun Hyundai field testing, and Apptronik has operational fleets at Robot Parks and partner sites. None has released a comparable mix of hours, cycles and completed units. For a warehouse buyer choosing a robot now, Digit is the safest bet based on what we can verify.

Has Figure 03 crossed from prototype to real product?

Yes. Figure 03 has crossed into repeatable production, even though much of its fleet is still feeding development rather than customer revenue.

Figure says BotQ has delivered more than 350 third-generation robots. The factory improved its demonstrated rate from one robot per day to one per hour in under 120 days. That is enough volume to create separate fleets for AI training, reliability testing, internal household work and customer projects.

The production detail gives the figure more weight. Figure reported more than 9,000 actuators built across several configurations, over 500 battery packs and a 99.3% first-pass yield on its battery line. Complete robots passed more than 80 functional tests before release. These numbers show a company dealing with repeatability, inspection, component variation and fleet maintenance.

BMW supplied the harder lesson. Figure 02's forearm became its most common hardware failure point during the 1,250-hour deployment. Figure redesigned the wrist electronics for Figure 03, removed a distribution board and simplified the cabling. That kind of redesign usually appears only after a machine has worked long enough to fail in predictable ways.

Figure 03 has now returned to BMW for sequencing, a workflow that combines thin-part handling with pulling a heavy wheeled cart. Figure also signed a commercial agreement to deploy humanoids in Catalyst Brands' distribution network. We still need customer throughput from the new generation, but the hand-built prototype phase is over.

If you want more recent data on this point, please see our latest humanoid robotics market report.

Can Boston Dynamics Atlas take the lead soon?

Atlas can take the overall lead soon, but its industrial work record is still too thin to award it first place.

Boston Dynamics unveiled a production-ready version of Atlas and began manufacturing it at its Boston headquarters. The first deployment slots were fully committed to Hyundai and Google DeepMind, with Hyundai using the robot for automotive sequencing work. The company also redesigned Atlas around fewer unique parts and automotive supply chains, while Hyundai Mobis agreed to supply actuators.

The product already includes features many rivals still describe as future goals. Atlas can work autonomously, through teleoperation or with tablet steering. Orbit software connects it to warehouse and manufacturing systems, distributes learned skills across a fleet and tracks performance. Self-swapping batteries address one of the dullest but most important barriers to continuous work.

What we lack is a public workload. Boston Dynamics has yet to release an Atlas equivalent to Digit's tote count or Figure's BMW runtime. The production Atlas has appeared in high-profile demonstrations and field testing, yet those events say little about intervention rates, cycle time or maintenance after months of use.

Atlas can do more physically and has a credible route to stronger intelligence through Google DeepMind. One sustained customer deployment could change the answer quickly. For now, Figure has simply shown more.

Is Tesla Optimus the most advanced humanoid robot?

Optimus currently trails the leaders because Tesla has published a factory plan without comparable autonomy or deployment results.

Tesla has several advantages that would be hard for a startup to copy. It already designs motors, power electronics, batteries, AI chips, vision systems and high-volume factories. Its own plants offer an enormous internal testing ground, and the company can fund a long development cycle without depending on one outside customer.

Its latest detailed manufacturing disclosure said Optimus Gen 3 would be the first version designed for mass production. Tesla was preparing the first California production line, with a planned start before the end of 2026 and an eventual annual capacity of one million robots. A later company update still described the program as progressing ahead of mass production.

Capacity targets deserve attention, but they tell us nothing about how often Optimus completes an unfamiliar task, how much human assistance it needs or how long the hardware runs between failures. Tesla has published few task-success rates, customer deployments or operating totals that we can compare with Figure and Agility.

Optimus may become the fastest-moving challenger once hundreds or thousands of machines begin collecting factory data. Today, Tesla's strongest evidence concerns the organisation it can build around the robot. The robot itself still falls short of first place.

Are Chinese humanoid robot companies already ahead?

China leads humanoid affordability and manufacturing depth, while Figure and Agility still hold stronger public evidence on broad autonomy and sustained work.

Unitree has changed who can participate in humanoid development. Its G1 sells from about $13,500 through the company's store, and its newer H2 offers a larger, stronger platform. NVIDIA recently chose the H2 Plus, fitted with Sharpa five-fingered hands and Jetson Thor compute, as the body for an open Isaac GR00T reference humanoid. That gives researchers a ready-made route from data collection to training and onboard control.

UBTECH is further along in industrial rollout. Walker S2 can reportedly swap its own battery in three minutes, handle 15-kilogram loads across a workspace running from the floor to 1.8 metres and coordinate through the company's factory software. UBTECH says the robot is in mass production and delivery, and earlier Walker models have trained in automotive plants including BYD.

China's wider advantage sits around the robot. The International Federation of Robotics reports that the country already operates about two million industrial robots and installs more than half of the world's new units. That supplier base, manufacturing experience and domestic demand can shorten hardware cycles and lower component costs.

The missing piece is transparent, comparable work data. Chinese companies publish many factory announcements and impressive motion videos, but fewer long autonomous sequences or six-figure customer throughput results. China may already lead the industrial ecosystem. The best all-around robot still comes from the United States.

If you want more recent data on this point, please see our latest humanoid robotics market report.

Does 1X have the best humanoid robot for the home?

NEO has the clearest home humanoid product, but Figure has shown more autonomous household ability.

1X has designed around the risks and inconveniences of a normal house. NEO uses a soft exterior, covered joints, tendon-driven movement and self-charging. Customers can reserve a $20,000 ownership version or choose a $499 monthly subscription. The company says full-scale production has started inside a 58,000-square-foot factory with more than 200 employees.

The newly announced hands make the product more serious. They combine tactile sensing with backdrivable tendons and unusually human-like movement. That should help with fragile objects and accidental contact, two problems that matter far more around people, pets and furniture than on a fenced factory line.

Autonomy remains the weak point. 1X openly says early NEO units arrive with basic autonomy. When the robot meets an unfamiliar task, a company expert can remotely supervise it at a scheduled time. That approach may collect valuable data quickly, but the customer is partly buying access to a human-supported learning service.

Figure has demonstrated richer independent household behaviour, including kitchen cleanup, laundry, room tidying and two-robot bed making. Figure has yet to package that ability into a consumer offer with NEO's safety design, price and support model. NEO leads as a home product; Figure leads the underlying household autonomy race.

Is Apptronik the sleeper in the humanoid robot race?

Apptronik is the most credible sleeper, with Apollo 2 serving mainly as a learning platform before the commercial Apollo 3 arrives.

Apollo 2 comes in both walking and wheeled versions. That choice is practical. Wheels suit flat warehouses, use less energy and fit existing industrial safety rules more easily. Legs remain available for sites where steps, tight passages or human-designed layouts justify the extra complexity.

Apptronik recently expanded its Austin Robot Park to nearly 90,000 square feet. Apollo 2 fleets also operate at Google DeepMind, Mercedes-Benz, GXO and other partner sites. The robots collect data through a mix of teleoperation and autonomous execution across logistics, manufacturing and retail tasks. That data feeds work on Google's Gemini Robotics models.

The company has also raised more than $935 million in its expanded Series A, with backing from Google, Mercedes-Benz and John Deere. Few humanoid startups can match that combination of robotics history, industrial customers, capital and access to a leading AI lab.

Apptronik's own description explains why it remains below the top tier. Apollo 2 is the workhorse for data collection, and the company says those lessons will power Apollo 3. Public results still lack the hours, success rates and throughput needed to rank Apollo beside Digit or Figure. The setup is impressive. The proof is still missing.

Are any humanoid robots genuinely general-purpose today?

None of today's humanoids is genuinely general-purpose. The best systems can stretch one architecture across a growing set of trained tasks.

Figure's Helix can reuse the same broad learning approach across logistics, kitchens, bedrooms and factory work. Each new capability still needs relevant data, training and a controlled rollout. Digit can switch between several warehouse workflows, yet it cannot walk into an unfamiliar plant and decide how to operate every machine. Atlas learns new industrial tasks quickly, but customers still go through application planning and training.

NEO makes the boundary especially clear. The product handles basic learned chores and calls on remote experts for harder ones. Apptronik gathers teleoperated data to prepare future autonomous fleets. Tesla continues building the data and hardware systems required for a production model.

Humans bring years of touch, movement and common sense to a new room. These robots start with much less physical experience and must avoid damaging expensive hardware while learning. Their apparent intelligence can also collapse when lighting, object shape, friction or timing changes beyond the training range.

The useful term is multi-purpose. A strong humanoid can learn several unrelated jobs on one body and carry parts of that learning into the next job. That is already valuable. General-purpose autonomy, where a robot safely handles a genuinely unfamiliar request without preparation, remains unsolved.

Who leads each part of the humanoid robot race?

The category leaders split cleanly: Atlas owns hardware, Digit owns workplace proof and Figure owns the best overall blend.

Figure wins our combined ranking because it stays near the front across almost every important test. Its body is capable enough for factory and household work. Helix 02 has the broadest learned task record, Figure 02 supplied real deployment data, and BotQ has produced hundreds of the new generation.

Atlas loses points mainly through missing customer throughput. Its hardware lead is large enough that one strong industrial deployment could move it into first place. Digit faces the opposite problem. Its work record is excellent, while its manipulation range and general task evidence remain narrower.

The rest of the field leads specific bets. Tesla has the best chance to reach huge manufacturing scale. Unitree offers the best accessible hardware ecosystem. UBTECH has the best Chinese industrial package. NEO is closest to a product built specifically for ordinary homes. Apptronik has assembled perhaps the best ingredients for a future jump.

Meaning of "most advanced"	Current leader	Why we chose it
Best overall humanoid	Figure 03	Strongest mix of AI, manipulation, deployment learning and production
Best physical hardware	Boston Dynamics Atlas	Highest strength, reach, joint freedom and environmental protection
Most proven worker	Agility Robotics Digit	Clearest sustained commercial throughput
Best demonstrated autonomy	Figure 03 with Helix 02	Broad, long, full-body tasks across several settings
Best demonstrated manipulation	Figure 03	Widest useful mix of rigid, flexible and two-handed object work
Best affordable platform	Unitree	Capable humanoids sold openly at unusually low prices
Biggest scale potential	Tesla Optimus	Manufacturing assets and a planned high-volume line
Strongest Chinese industrial robot	UBTECH Walker S2	Production and broad factory access
Strongest dark horse	Apptronik Apollo	DeepMind partnership, Robot Parks and strong industrial backing

Who has the most advanced humanoid robot today?

Figure currently has the most advanced humanoid robot overall: Figure 03 powered by Helix 02.

The answer rests on the combined package. Helix 02 has shown full-body autonomy across kitchens, living rooms, bedrooms, logistics and automotive work. Figure's previous generation accumulated 1,250 hours at BMW and loaded more than 90,000 parts. The company has since produced over 350 Figure 03 units, returned the new model to BMW and signed another logistics customer.

Boston Dynamics comes closest. Atlas has the best body by a comfortable margin and now looks like serious industrial equipment. Its 50-kilogram peak lift, 2.3-metre reach, continuous joints, IP67 protection and self-swapping battery let it do more physically than Figure. We need months of quantified customer work before moving it to the top.

Agility deserves the clearest commercial credit. Digit's 100,000 tote movements show that a humanoid can already become part of an ordinary warehouse process. Its narrower manipulation and AI record keep it behind Figure in an overall technology ranking.

Tesla, Unitree, UBTECH, 1X and Apptronik could all change the answer. Their strongest claims concern scale, affordability, factories, home design or future AI rather than a fully rounded, proven machine.

Figure 03 leads because it combines the best public autonomy record with useful hands, whole-body control, real factory learning and a fleet large enough to improve quickly. Atlas has the body to overtake it, and Digit has the work record Figure still needs to match. Today, Figure has the best all-around humanoid robot.

If you want more recent data on this point, please see our latest humanoid robotics market report.

Who is the author of this content?

NEW MARKET PITCH TEAM

We track new markets so founders and investors can move faster

We build living "market pitch" documents for emerging markets: AI, synthetic biology, new proteins, and more. Instead of outdated PDFs or hallucinated LLM answers, our clients get a clean, visual, always-updated view of what's really happening: key players, deals, regulations, and signals that matter. Learn more about us.