Tesla Optimus Gen 3: version history

Current record →

Nothing is deleted from the index. Every change to this record is kept: a full copy of each version since 6 Sep 2026, 05:24 UTC, and the field-by-field updates applied before that. See every robot from Tesla.

Versions · 7 recorded

Version 545 · 28 Oct 2026, 16:56 UTC · score then 65.2
- ai model: — → Grok (xAI) for voice interaction; Tesla end-to-end neural network stack
- battery (kWh): — → 2.3
- deployments: +1 added
- field evidence: — → 9 fields
- verified fields: +1 added
Changed by: Update Optimus Gen 3 (reviewer)
Version 519 · 11 Oct 2026, 23:26 UTC · score then 64.5
- dexterous hands: — → True
- hand dof per hand: — → 22
- onboard compute: — → Tesla AI5 inference chip
- verified fields: +4 added
Changed by a direct correction or a scoring update, not a research proposal.
Version 507 · 3 Oct 2026, 05:36 UTC · score then 45.4
- ai model: Tesla FSD end-to-end neural net → —
- battery (kWh): 2.3 → —
- confidence: medium → low
- deployments: +1 added, 2 removed
- dexterous hands: True → —
- first unveiled year: 2026 → —
- hand dof per hand: 22 → —
- height (cm): 173 → —
- notable: Optimus Gen 3 enters pilot production in 2026 with 22-DoF hands, AI5 chip, and Grok integration. It is being tested in 24/7 autonomous shifts at Tesla factor… → Tesla's third Optimus generation (V3). Tesla builds it on the first-generation line at Fremont, which replaced the Model S and X lines; Electrek reported sev…
- onboard compute: Tesla HW4.0 → —
- payload (kg): 20 → —
- price (USD): 20000 → —
- runtime (h): 10.0 → —
- sources: +3 added, 2 removed
- specs: 8 fields → —
- speed (m/s): 1.2 → —
- total dof: 81 → —
- verified fields: +2 added
- weight (kg): 57 → —
Changed by: Update Optimus Gen 3 (confirmed: onboard_compute, hand_dof_per_hand, dexterous_hands, form_factor) (published automatically)
Version 501 · 3 Oct 2026, 04:24 UTC · score then 81.6
- ai model: VLA (Vision-Language-Action) model with Grok voice integration → Tesla FSD end-to-end neural net
- onboard compute: Tesla AI5 (Hardware 5) custom inference processor → Tesla HW4.0
- price (USD): — → 20000
- runtime (h): 9 → 10.0
- speed (m/s): — → 1.2
Changed by: Update Optimus Gen 3 (confirmed: ai_model, onboard_compute, speed_ms, runtime_hours, price_usd) (published automatically)
Version 498 · 25 Sep 2026, 04:50 UTC · score then 85.1
- onboard compute: AI5 chip → Tesla AI5 (Hardware 5) custom inference processor
Changed by a direct correction or a scoring update, not a research proposal.
Version 333 · 15 Sep 2026, 04:28 UTC · score then 85.1

No field changed; the score was recalculated.

Changed by: Update Optimus Gen 3 (confirmed: onboard_compute) (published automatically)
First recorded version · 6 Sep 2026, 05:24 UTC · score then 85.1

The record as it stood when full copies began to be kept.

Earlier changes · 4, before full copies were kept

Update Optimus Gen 3 · 2026-09-04 23:10 UTC · published automatically
- onboard compute: Tesla AI5 chip (taped out April 2026) → AI5 chip
Update Optimus Gen 3 · 2026-09-03 22:54 UTC · published automatically
- ai model: Grok → VLA (Vision-Language-Action) model with Grok voice integration
- onboard compute: AI5 → Tesla AI5 chip (taped out April 2026)
- total dof: 50 → 81
Update Optimus Gen 3 · 2026-09-01 22:25 UTC · published automatically
- status: pilot → limited-production
- ai model: End-to-end vision neural networks with xAI Grok integration → Grok
- onboard compute: Tesla AI5 inference chip → AI5
- runtime (h): — → 9
- battery (kWh): — → 2.3
- total dof: 72 → 50
- deployments: updated
- first unveiled year: 2024 → 2026
New model: Optimus Gen 3 (Tesla) · 2026-08-23 01:04 UTC · research agent

Record created.