Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

@manikv12
Copy link
Contributor

Summary

Add hf:moonshotai/Kimi-K2.5 to the synthetic model catalog.

Changes

  • Added Kimi K2.5 model to SYNTHETIC_MODEL_CATALOG in src/agents/synthetic-models.ts
  • Model specs: 256k context window, 8192 max tokens, supports reasoning
  • Available via https://dev.synthetic.new/docs/api/models

Testing

  • Verified model appears in openclaw models list --all --provider synthetic
  • Model ID: synthetic/hf:moonshotai/Kimi-K2.5

Copilot AI review requested due to automatic review settings January 30, 2026 05:45
@moltbot-barnacle moltbot-barnacle bot added the agents Agent runtime and tooling label Jan 30, 2026
@moltbot-barnacle moltbot-barnacle bot added the commands Command implementations label Jan 30, 2026
Copy link

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Adds the hf:moonshotai/Kimi-K2.5 model to the Synthetic provider’s static model catalog so it shows up in model listing and selection flows.

Changes:

  • Added a new SYNTHETIC_MODEL_CATALOG entry for hf:moonshotai/Kimi-K2.5
  • Declared model capabilities/specs (context window, max tokens, input types, reasoning flag)

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment on lines +104 to +106
name: "Kimi K2.5",
reasoning: true,
input: ["text"],
Copy link

Copilot AI Jan 30, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

reasoning is set to true for hf:moonshotai/Kimi-K2.5, but elsewhere in the codebase Kimi K2.5 is treated as non-reasoning (e.g., Moonshot provider config sets reasoning: false for Kimi K2.5 in src/agents/models-config.providers.ts:288-295). If this model does not actually support the app’s reasoning mode, this flag will incorrectly enable reasoning defaults and directives. Please align the flag with the established Kimi K2.5 capability (likely false), or add a clear rationale if Synthetic’s hosted variant behaves differently.

Copilot uses AI. Check for mistakes.
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@copilot open a new pull request to apply changes based on this feedback

manikv12 and others added 3 commits January 30, 2026 07:02
Add hf:moonshotai/Kimi-K2.5 to the synthetic model catalog.
This model is available via dev.synthetic.new API.

- 256k context window
- 8192 max tokens
- Supports reasoning
@steipete steipete force-pushed the add-kimi-k2.5-model branch from 231c9eb to 6bd9823 Compare January 30, 2026 06:17
@steipete steipete merged commit 5e635c9 into openclaw:main Jan 30, 2026
18 of 23 checks passed
@steipete
Copy link
Contributor

Landed via temp rebase onto main.

  • Gate: pnpm lint && pnpm build && pnpm test
  • Land commit: 6bd9823
  • Merge commit: 5e635c9

Thanks @manikv12!

@manikv12 manikv12 deleted the add-kimi-k2.5-model branch January 30, 2026 06:33
manikv12 added a commit to manikv12/openclaw that referenced this pull request Jan 30, 2026
…definitions

Kimi K2.5 supports vision but was declared text-only in both the synthetic
catalog and the moonshot provider.  More importantly, onboarding bakes the
full model catalog into openclaw.json, and the old merge logic let stale
config definitions shadow code updates — so even after fixing the catalog,
existing users would never see the change.

Fix mergeProviderModels so code-defined capability fields (input, reasoning,
contextWindow, maxTokens) override stale config entries while preserving
user-specific fields (cost, headers, compat, apiKey).

Follow-up to 5e635c9 (openclaw#4407).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

agents Agent runtime and tooling commands Command implementations

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants