PulseAugur
EN
LIVE 10:03:41

Meituan releases LongCat-2.5 with 1.6T parameters for long-horizon agents

By PulseAugur Editorial · [1 sources] ·

Meituan has released LongCat-2.5, a large language model with 1.6 trillion parameters, approximately 48 billion of which are active. This natively multimodal model is designed for long-horizon agents and was trained entirely on Chinese hardware. It features a unique approach to image quantization and incorporates reinforcement learning for failed trajectories, achieving 43% faster prefill inference speeds. AI

IMPACT This model's advanced multimodal capabilities and efficiency could push the boundaries for long-horizon agent development.

RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Meituan releases LongCat-2.5 with 1.6T parameters for long-horizon agents

How we ranked this

Signal score
4 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Same-day
Cluster formed today. Ranking reflects the current source set at time of score.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · Yi ·

    Meituan shipped LongCat-2.5

    <p>Meituan shipped LongCat-2.5. 1.6T params, ~48B active, native multimodal, aimed at long-horizon agents, trained start to finish on Chinese hardware.</p> <p>Images get quantized into multi-level discrete tokens, and that's where Visual Coding and 3D pick up. On the RL side, fai…

RELATED ENTITIES

RELATED TOPICS