Meituan releases LongCat-2.0: 1.6T open-source coding model on Chinese chips
Meituan has open-sourced LongCat-2.0, a 1.6 trillion-parameter Mixture-of-Experts (MoE) coding model built for agentic software engineering. LongCat-2.0 activates about 33B–56B parameters per token rather than using all parameters at once, improving efficiency. It supports a 1 million-token context window for large codebase reasoning.
On benchmarks, LongCat-2.0 scored 59.5 on SWE-bench Pro and 70.8 on Terminal-Bench. Model weights and resources are publicly available on Hugging Face under the “meituan-longcat” organization.
Crucially, Meituan says LongCat-2.0 was trained and run on a domestic 50,000-card compute cluster using Chinese-manufactured hardware, with no reliance on restricted top-end AI chips such as Nvidia A100/H100 or AMD MI300X. Meituan frames this as a milestone for China’s local compute at this scale.
LongCat’s prior releases include LongCat-Flash (560B) in Sep 2025 and LongCat-Next (multimodal) in Mar 2026; LongCat-2.0 arrives on Jun 30, 2026, nearly tripling parameters in under a year. For developers, open availability lowers barriers to fine-tuning for tasks like code generation, security auditing, and automated debugging.
Neutral
This news is about an open-source AI model release (LongCat-2.0) and China’s domestic compute capabilities, with no direct connection to specific crypto assets, on-chain flows, or protocol changes. As a result, it is unlikely to trigger immediate, measurable trading effects in major coins.
That said, frontier-model and compute self-sufficiency narratives can sometimes influence sentiment toward AI- and tech-adjacent sectors. In past crypto cycles, broad “tech progress” headlines have occasionally caused short-lived risk-on moves in thematic tokens, but without token-specific catalysts (e.g., listings, partnerships, protocol upgrades), the impact usually fades quickly and remains mostly sentiment-driven.
Short-term: neutral. Traders are more likely to treat this as general tech/AI news rather than a catalyst for BTC/ETH volatility.
Long-term: neutral to mildly positive for AI-infra narratives, but any crypto market relevance would depend on whether this translates into ecosystem funding, tokenized AI infrastructure demand, or major corporate involvement—none of which is stated here.