Buying guide
Omi vs RayNeo V3 in 2026
Comparing 2 humanoid robots across availability, pricing, capabilities, and verified deployments. Current as of 2026.
Key differences
Neither Omi nor RayNeo V3 leads across the verified dimensions on DEPLOY's record as of 2026; the comparison table holds every sourced figure side by side.
- Omi has the lower recorded price.
| Attribute | No image on file | |
|---|---|---|
| Manufacturer | Based Hardware | RayNeo |
| Form factor | wearable | wearable |
| Maturity | commercial | commercial |
| Autonomy | — | — |
| Availability | shipping-nowshipping | shipping-nowshipping |
| Price | $17-$19/month subscription or $89 (actual sale price) | $245-$255 (actual sale price) |
| Capability claims | — | — |
| Brain | — | — |
| Verified deployments | 1Based Hardware | 1RayNeo |
| Privacy practices | 5third-party-sharing, on-device-processing, capture-indicator, data-deletion-control, cloud-upload | 4data-sale, third-party-sharing, data-retention, data-deletion-control |
| Sources on file | 6 | 5 |
Track this matchup
Get told the moment Omi or RayNeo V3 changes. Verified changes only, the second they are real. No press-release noise.
Omi
RayNeo V3
Editorial summaries
Omi
Omi, by Based Hardware (founder Nik Shevchenko, San Francisco), is a shipping AI necklace in the ambient-recording 'second brain' category: a small circular orb on a neck lanyard that captures conversations and runs them through GPT-4o for ambient transcription in 25+ languages, auto-summaries and tasks, a searchable memory database, daily recaps, and a 'Brain Map'. It sells for $89 as a consumer unit (about $70 dev) one-time, with a free tier (unlimited on-device transcription plus 1,200 cloud minutes a month) and an optional Omi Unlimited plan from $16 a month; the user owns the device.
It is one of the few genuinely open-hardware entries in the cohort: an MIT-licensed GitHub repo of about 12,700 stars with firmware, apps, and open hardware designs, maintained into 2026. The AI is genuine cloud functionality, not veneer. One significant cap-flag: the aspirational brain-interface and EEG 'reads-your-mind' marketing is not the shipping product, which is audio-only; the BCI module is roadmap with no code, and TechCrunch could not verify it. (Omi is distinct from Friend, Avi Schiffmann's pendant; Shevchenko's device was originally also named Friend and was renamed Omi.)
RayNeo V3
The RayNeo V3 (from RayNeo, TCL's AR glasses subsidiary) is a camera-and-audio AI glasses product with no display, in the Ray-Ban Meta class: 39 grams, a 12-megapixel TCL co-developed camera (1080p/30fps), and 32GB of storage. It went on sale in China in January 2025 at 1,799 yuan, roughly $249.
Its AI is genuine: a customized multimodal large language model via an Alibaba Tongyi Qianwen partnership, with three years of free AI updates, placing it in the genuine-cloud-AI tier. Two cap-flags: pricing is China-market CNY (the USD figure is a conversion, and it is not officially sold outside China), and the vendor's performance numbers (a 1.3-second response time and 98 percent accuracy) are RayNeo-stated, not independently verified. It is the TCL/RayNeo parallel to Xiaomi AI Glasses and Baidu's Xiaodu in the Chinese consumer wearable-AI cohort.
Common questions
How do Omi and RayNeo V3 differ?
What is the difference between Omi and RayNeo V3?
Which is cheaper, Omi or RayNeo V3?
Recent coverage
- $IOTEX target for bullrunOmi · TradingView · 2026-08-22
- NYT Pips hints, answers for August 22, 2026Omi · Mashable · 2026-08-22
- Boom! Hundreds of New Products Launched Frenetically, Billions of Funds Pouring In: Ten Founders Summarize 2025 Smart Glasses History in Tens of Thousands of Words - 36 KrRayNeo V3 · Google News · 2026-01-06
- China’s Telecom Giants Back Smart-Glasses Maker RayNeo in $143 Million Funding Round - Caixin GlobalRayNeo V3 · Google News · 2026-01-06
Related comparisons
Keep researching
Keep reading
Ask the registry
For AI assistants
Use DEPLOY in Claude, Cursor, and other AI assistants
Connect any MCP-aware assistant to the live registry. Ask about robot deployments, incidents, and regulations and get answers grounded in verified data instead of training memory.
Machine-readable: this page as markdown.
