aifollow.news Search
Back IT之家 科技新闻
IT之家 科技新闻· · Original publication time AI score40

Momenta previews R7 Ultra world model, claiming fivefold parameter scaleMachine translation

Momenta 预告 R7 Ultra 世界模型,参数规模称将扩大至 5 倍

Automatically verified and published · Generated and evidence-checked automatically; not reviewed by a human.

AI-assisted summary

Momenta announced at an XHEART event that it will release the R7 Ultra world model, claiming five times the previous parameter scale and a fivefold improvement in safety performance. The model is intended to run in vehicles using the X7 Ultra chip. Momenta says it can handle complex driving scenarios including U-turns, passing encounters and parking; these capabilities remain the company's claims.

Article · AI translation

IT Home reported from the scene on October 29 that Momenta announced at its strategic partner XHEART’s New Chip Journey launch event that it will soon release the R7 Ultra world model.

According to the introduction, drawing on accumulated data and XHEART’s X7 Ultra large-model chip, the Momenta R7 Ultra model has expanded its parameter count to 5 times the original and improved its safety performance by 5 times, in what the company calls a shift “from understanding the world to simulating how it unfolds.”

IT Home learned at the launch event that XHEART’s New Chip Journey X7 Ultra provides 700 TOPS of computing power for large models. Combined with an ultra-large-core NPU architecture, it supports running larger models in vehicles, giving R7 Ultra more intelligent capabilities for understanding, simulating and predicting scenarios.

Momenta said that in complex situations such as making U-turns across multiple lanes, negotiating with oncoming vehicles, reversing to yield and parking in extremely tight spaces, R7 Ultra can demonstrate driving ability comparable to, or even better than, that of experienced human drivers, making driving smoother and more natural. In the future, it plans to further expand the model’s parameter count to 10 times the original and improve safety performance by more than 10 times.

Found an error? Send a correction