Meituan tests trillion-parameter AI model built on domestic compute

By PulseAugur Editorial · Summary by gemini-2.5-flash-lite from 1 source

Meituan has reportedly initiated a private test of a trillion-parameter AI model, developed using only Chinese computing infrastructure. This model is said to rival GPT-4's performance and was likely trained using Huawei's Ascend hardware, bypassing NVIDIA components. The development signifies Meituan's strategic move into foundational AI model development. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Signals increased competition in foundational model development from Chinese tech giants.

RANK_REASON Trillion-parameter model release from a major tech company, comparable to GPT-4.

Read on Mastodon — mastodon.social →

COVERAGE [1]

Mastodon — mastodon.social TIER_1 · [email protected] · 2026-04-27 04:34

Meituan has quietly launched an invitation-only test of a trillion-parameter AI model trained entirely on domestic computing clusters, according to sources. The

Meituan has quietly launched an invitation-only test of a trillion-parameter AI model trained entirely on domestic computing clusters, according to sources. The model is broadly comparable to GPT-4 in capability and is believed to have been trained without NVIDIA hardware, likely…

COVERAGE [1]

Meituan has quietly launched an invitation-only test of a trillion-parameter AI model trained entirely on domestic computing clusters, according to sources. The

RELATED ENTITIES

RELATED TOPICS