モデルを比較

This is the batch-processing variant of Muse Glimmer 30B, Meta Superintelligence Labs' dense, open-weight 30B-parameter multimodal model (Apache 2.0 license, 131K-token context) distilled from Muse Spark and optimized to run agentic workloads locally on a single consumer GPU, released August 2026. The batch endpoint processes requests asynchronously at a lower per-token cost, well suited for large-scale offline tool-use, coding, or long-horizon agent evaluation rather than latency-sensitive interactive use.

Claude Haiku 4.5 is Anthropic's fast, affordable small model in the Claude 4 generation, balancing strong instruction-following and coding ability with low latency. It is designed for high-throughput agentic and chat applications where speed and cost efficiency are priorities.