An Honest Frame for the Comparison
These two robots are not at the same stage of maturity, so a spec-table comparison would mislead. The Unitree G1 is a product you can buy, that ships, and that is running in labs today. Tesla Optimus is a development programme being trialled in Tesla's own facilities, with no open sale to end customers. Below we give the G1's verifiable numbers; on the Optimus side we speak only to the publicly known status of the programme and do not write estimated technical figures.
The Unitree G1's Verifiable Data
1320 × 450 × 200 mm standing at 35 kg, folding to 690 × 450 × 300 mm. 23 degrees of freedom, 2 kg arm payload (3 kg on the EDU) and a top speed of 2 m/s (7.2 km/h). Around 2 hours from a quick-release 9000 mAh lithium-ion battery, charged at 54V 5A. On compute, a 100 TOPS class Jetson Orin platform. The full list and configuration options are on the G1 product page.
Commercial Availability
In practice this is the largest difference. The G1 can be ordered in Turkey through Robotlar.org, with customs, VAT, installation, training and warranty all defined. See our G1 pricing page for current configurations. For a team that wants to start working with a humanoid this year, the comparison ends here.
Developer Access and Ecosystem
The G1 EDU grants full SDK and ROS 2 access; low-level joint control, running your own navigation stack and deploying a reinforcement learning policy all depend on it. NVIDIA's Isaac GR00T platform also supports the G1, so a reference workflow from simulation to the real robot can be built; our GR00T guide walks through it step by step.
Who Should Choose What
A university lab, an R&D team, a thesis project, an organisation that wants to start working with a humanoid this year: the G1, and the EDU if development is involved. Following Optimus makes sense, but it is not a line item you can put into a budget or a project schedule today. If you need a full-size platform on the Unitree side, turning to the H2 rather than Optimus makes for a more realistic comparison.