Skip to content

fix: [privatemode] sync supported models - #6014

Open
github-actions[bot] wants to merge 1 commit into
devfrom
issue-6013
Open

fix: [privatemode] sync supported models#6014
github-actions[bot] wants to merge 1 commit into
devfrom
issue-6013

Conversation

@github-actions

@github-actions github-actions Bot commented Sep 1, 2026

Copy link
Copy Markdown
Contributor

Summary

Sync Privatemode AI (providers/privatemode-ai) with public docs as of v1.55.0. The supported-model overview adds GLM-5.3 (glm-5.3) and the glm-latest alias; those entries were missing from the catalog. Existing Privatemode model TOMLs already matched the current overview and EUR price list, so they were left unchanged.

Changes

Added

  • providers/privatemode-ai/models/glm-5.3.toml

    • base_model = "zhipuai/glm-5.3"
    • status = "beta" (docs mark it preview)
    • reasoning_options = [{ type = "effort", values = ["low", "high", "max"] }] (always-on reasoning; none maps to max on this host)
    • limit.context = 256_000 (host cap; lab base is larger)
    • Cost (USD/MTok from EUR list): input 1.7910, output 8.9436, cache_read 0.1733
  • providers/privatemode-ai/models/glm-latest.toml

    • Same backing model, pricing, reasoning, and context as glm-5.3
    • name = "GLM (latest)" (alias that tracks the latest GLM on Privatemode)

Unchanged (already aligned with docs)

  • deepseek-ocr-2, gpt-oss-120b, kimi-k2.6, kimi-latest, qwen3-embedding-4b, voxtral-mini-3b, whisper-large-v3

Not added

  • glm-5.2: deprecated; routes to GLM-5.3 and is not in the overview table’s desired ID set

Evidence

Claim Source What it establishes
Supported IDs include glm-5.3 and glm-latest; GLM-5.3 is preview; context 256k; text input Privatemode models overview Desired catalog + host context/modalities
Effort low / high / max (default max); nonemax; reasoning always on Overview § GLM-5.3 reasoning_options
v1.55.0 upgrades to GLM-5.3; glm-5.2 deprecated Release notes v1.55.0 Why add GLM-5.3 / skip standalone 5.2
Upstream identity zai-org/GLM-5.3 Hugging Face zai-org/GLM-5.3 Lab inheritance target
Lab metadata already present models/zhipuai/glm-5.3.toml Safe base_model without editing models/
EUR prices: 1.55 / 7.74 / 0.15 cached input per MTok Privatemode pricing Cost inputs before USD conversion
FX 1 EUR = 1.1555 USD (Frankfurter 2026-08-10) Existing Privatemode TOMLs (e.g. kimi-k2.6.toml) Same EUR list → same USD figures as Kimi on this provider

Validation

  • Compared the full overview table to providers/privatemode-ai/models/*.toml
  • Verified GLM pricing, context, preview status, and reasoning controls against Privatemode docs
  • Confirmed zhipuai/glm-5.3 lab entry exists for base_model
  • Did not run bun validate (no shell/Bash in this workflow)

Review notes

  • USD costs use the provider’s established 1.1555 EUR→USD rate (same as Kimi’s identical EUR list on this host), not a fresh FX rebase of the whole Privatemode catalog.
  • glm-5.2 omitted on purpose (deprecated, not in the overview table).
  • Canonical models/ files were not modified, per issue scope.

Closes #6013

Automated by the issue fixer: https://github.com/anomalyco/models.dev/actions/runs/33537104665

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[privatemode] sync supported models

0 participants