v0.40.16
In this release, Ripple Preview debuts as a 65B dense model purpose-built for agentic coding, and Qwen3.8 Flash Next trims about 27 GiB of memory on Apple silicon.
Nov 10, 07:45
Why this: Covers model and Agent · Relevant to this topic
2 items · All topics →
In this release, Ripple Preview debuts as a 65B dense model purpose-built for agentic coding, and Qwen3.8 Flash Next trims about 27 GiB of memory on Apple silicon.
Why this: Covers model and Agent · Relevant to this topic
This release adds support for Xiaomi's 9B MiMo V2.6 model, built for code and agent workflows, plus AutoTrust decision models that read text and make judgments.
Why this: Covers model and Agent · Relevant to this topic