TurboVLA Runs at 32Hz on RTX 4090 with <1GB VRAM, Outperforming Larger Models
jiqizhixin · x · 2026-08-13
Huazhong University of Science and Technology and Huawei jointly introduced TurboVLA, a model designed to solve the high computational cost of traditional robots relying on massive language models.
Instead of routing everything through a heavy LLM, TurboVLA allows vision and language to interact directly and predicts robot actions in real-time, bypassing the LLM middleman and drastically reducing compute and memory requirements.
- Performance: Achieves a 97.7% success rate on the LIBERO robotics benchmark.
- Hardware: Runs at 32 Hz on a consumer-grade RTX 4090 using less than 1GB of VRAM.
- Comparison: Outperforms much larger VLA models.
More from Embodied
- AutoAgent0: VLM-powered agentic runtime lifts long-tail driving success by 28.3pp — yining_hong · 2026-10-16
- CarryOn: dual-loop harness adds personalization and safe recovery to frozen VLA robots — Zixuan Wang · 2026-10-16
- FoveaVLA replaces wrist cameras with learned foveation, hitting 70.8% success — Haoyu_Xiong_ · 2026-10-16
- Musk teases two-way Grok x Optimus integration: Grok in the bot, bot in Grok — beffjezos · 2026-10-16
- SimVTA uses simulator-rendered video to post-train driving world action models with RL — OpenDriveLab · 2026-10-16
- Robotics data startups compared: Scale AI tops $1B revenue as 9 LLM vs robotics data suppliers benchmarked — FinanceYF5 · 2026-10-16