Article · Meta AI · en
Llama 3.1 made Meta's open-model strategy explicit
Published 23 JUL 2024 · Curated 22 JUL 2026
What the source reports
Meta released the 405-billion-parameter Llama 3.1 model in July 2024 alongside a broader ecosystem pitch covering clouds, inference providers, distillation, safety tools, and a proposed Llama Stack interface.
Editorial analysis
The release is both a technical artifact and interested platform strategy. Meta's capability, cost, and safety comparisons are first-party claims; its partner list and license changes are records that can be checked directly.
Why it matters
The strongest comparison with China's model ecosystem is not open versus closed by nationality, but how durable each firm's incentives are to keep weights, tools, and derivative rights available.
Claims and evidence
reported · Meta AI
Meta released the 405-billion-parameter Llama 3.1 model in July 2024 alongside a broader ecosystem pitch covering clouds, inference providers, distillation, safety tools, and a proposed Llama Stack interface.
- This entry does not treat the canonical publisher as independent confirmation of its own report.
Source ledger
canonical source
Meta AI · Introducing Llama 3.1: Our Most Capable Models to Date ↗Published 23 JUL 2024 · Retrieved 22 JUL 2026 · en · original-english
Primary artifact for the 405B release, license change, ecosystem partners, and Meta's disclosed evaluation claims.
Technology · Economy · llama-3-1 · open-weights · platform-strategy
Mark Zuckerberg · Meta · United States
Related dispatches
- Llama 2 established the American open-weight precedent
Meta AI · 18 JUL 2023
- Liang Wenfeng described DeepSeek's playbook before the shock
ChinaTalk · 27 NOV 2024