Announcing MCP Toolbox Java SDK v1.0: Agentic data access for the enterprise
cloud.google.com·6d ago
Cloud & Infrastructure
The New Stack·Sep 2·5 min read
Multiverse says its 438B model is fast enough for AI agents. The benchmarks tell a more complicated story.
TL;DR
Multiverse Computing has launched Quasar 438B, a 438-billion-parameter AI model aimed at coding and enterprise agents, claiming it is fast and efficient. However, benchmarks reveal a more nuanced performance landscape compared to leading models.
✦ Why It Matters
Evaluate Quasar 438B for your coding and enterprise AI needs to leverage its unique performance metrics.
Key Takeaways
- Consider Quasar for coding and enterprise AI applications.
- Assess the impact of model compression on performance.
- Monitor hardware requirements for optimal deployment.
- Explore the implications of proprietary access on development.
Full Summary
Multiverse Computing introduced Quasar 438B, a 438-billion-parameter AI model aimed at coding and enterprise applications. It scored 43 on the Artificial Analysis Intelligence Index and 69.3 on Terminal-Bench v2.1, with an output speed of 183 tokens per second.
While it outperforms some competitors, it still lags behind top models like Claude Opus 5, which scored 89.1. Quasar features a context window of one million tokens, allowing it to handle large codebases effectively.
However, Multiverse has not disclosed the specifics of its compression technology or the hardware requirements for running Quasar, raising questions about its practical deployment. The model is proprietary and only accessible via the Multiverse CompactifAI API, limiting developers' ability to evaluate its performance independently.