Highlights

Technical patch. You don't need to do anything on your end — this is a model-catalog update. If a specific model was bound to a bot, it keeps working the same way.

What's new

New model

• Grok build 0.1 (xAI) — available in the model picker alongside the rest of the Grok lineup.

Cheaper

A few models got cheaper following the OpenRouter price refresh:

• Ling 2.6 1T (InclusionAI) — input and output cost dropped by about 4x.
• Qwen 3.5 35B-A3B — a small input-price cut.
• Qwen 3.5 397B-A17B — the cached-token discount was removed (input and output prices unchanged).

If you use these models, the new pricing applies automatically from the next call.

What stayed the same

• Connections, bots, integrations, chats, billing — unchanged.
• The API and SDK are compatible with v1.2.0. No breaking changes.

Upgrading

If you use Worken through the UI, no action needed. If you run a self-hosted build, upgrade to worken@1.2.1 (npm).