fix(providers): refresh Aliyun and Kimi model catalogs

[wangfei010313]

Description

Refresh the built-in Kimi, Aliyun Token Plan, and Aliyun Coding Plan model
catalogs to match the current official model availability and metadata.

The existing catalogs contain retired Kimi model IDs, omit newly available
models, and include unsupported Aliyun Coding Plan IDs.

Type of Change

• Bug fix
• Static provider catalog maintenance
• Model capability metadata update

Components Affected

• Providers
• Model catalogs
• Context-window metadata
• Provider tests

Changes

Kimi

Replace the retired Kimi static catalog with the current model lineup:

• Add kimi-k3
• Add kimi-k2.7-code
• Add kimi-k2.7-code-highspeed
• Add kimi-k2.6
• Keep kimi-k2.5
• Remove kimi-k2-0905-preview
• Remove kimi-k2-0711-preview
• Remove kimi-k2-turbo-preview
• Remove kimi-k2-thinking
• Remove kimi-k2-thinking-turbo
• Make kimi-k3 the recommended Kimi model

Update the corresponding China and international capability baselines.
All five current Kimi models support image and video input.

Update model limits according to the official Kimi documentation:

• kimi-k3: 1M context and 128K maximum output
• Kimi K2.5, K2.6, and K2.7: 256K context
• Kimi K2.5, K2.6, and K2.7: 32K maximum output

kimi-k3 keeps its supported reasoning_effort configuration with
low, high, and max.

Kimi K2.5, K2.6, and K2.7 intentionally do not expose configurable agent
thinking metadata. Their thinking protocols cannot be represented safely by
the current OpenAI-compatible provider abstraction. In particular,
kimi-k2.7-code is always-thinking and does not support
reasoning_effort or disabling thinking.

Aliyun Token Plan

Update the shared China/international Token Plan static catalog:

• Add qwen3.8-max
• Add kimi-k2.7-code
• Keep the existing supported Token Plan models

qwen3.8-max includes image/video capability, 1M context, and 128K maximum
output metadata.

kimi-k2.7-code includes image/video capability, 256K context, and 32K
maximum output metadata.

China and international Token Plan providers continue to share the same
static catalog because their current official model tables are identical.

Do not add qwen3.8-max-preview. The preview ID requested in #6285 has been
replaced by the current GA model ID, qwen3.8-max.

No Token Plan model discovery behavior is introduced or changed by this PR.

Aliyun Coding Plan

Replace the stale Coding Plan entries with the exact model IDs documented for
the China and international Coding Plan endpoints:

• qwen3.7-plus
• qwen3.6-plus
• qwen3.5-plus
• glm-5
• glm-4.7
• MiniMax-M2.5
• kimi-k2.5
• qwen3-max-2026-01-23
• qwen3-coder-next
• qwen3-coder-plus

Remove glm-5.1 and glm-5.2, which are not included in the official
Coding Plan allowlist. Model IDs retain the exact documented casing.

Catalog Metadata

• Update the packaged catalog version to 2026.08.25
• Add context-window lookup entries for Qwen3.8 and Kimi K3
• Refresh multimodal capability baselines and documentation references

Tests

• Update Kimi activation tests to use the current catalog
• Update the Kimi discovery merge fixture to cover API, catalog, and merged
  discovery origins
• Update the recommended Kimi model snapshot
• Remove redundant assertions that duplicate mutable static catalog contents
• Update packaged catalog counts for the final catalog structure

Documentation

• https://help.aliyun.com/zh/model-studio/token-plan-team-overview
• https://help.aliyun.com/en/model-studio/token-plan-team-overview
• https://help.aliyun.com/zh/model-studio/coding-plan
• https://www.alibabacloud.com/help/en/model-studio/coding-plan
• https://help.aliyun.com/zh/model-studio/qwen3-8-max
• https://platform.kimi.com/docs/models
• https://platform.kimi.com/docs/guide/kimi-k2-7-code-quickstart
• https://platform.kimi.com/docs/api/models-overview

Related Issues and Pull Requests

This PR supersedes #4682 and #6293.

Closes #6285.

[github-actions]

Welcome to QwenPaw! 🐾

Hi @wangfei010313, this is your 36th Pull Request.

📋 About PR Template

To help maintainers review your PR faster, please make sure to include:

• ✅ Description - What this PR does and why
• ✅ Type of Change - Bug fix / Feature / Breaking change / Documentation / Refactoring
• ✅ Component(s) Affected - Core / Console / Channels / Skills / CLI / Documentation / Tests / CI/CD / Scripts
• ✅ Checklist:
  • Run and pass pre-commit run --all-files
  • Run and pass relevant tests (pytest or as applicable)
  • Update documentation if needed
• ✅ Testing - How to test these changes
• ✅ Local Verification Evidence:

  pre-commit run --all-files
  # paste summary result
  
  pytest
  # paste summary result

Complete PR information helps speed up the review process. You can edit the PR description to add these details.

🙌 Join Developer Community

Thanks so much for your contribution! We'd love to invite you to join the official QwenPaw developer group! You can find the Discord and DingTalk group links under the "Developer Community" section on our docs page:
https://qwenpaw.agentscope.io/docs/community

We truly appreciate your enthusiasm—and look forward to your future contributions! 😊

We'll review your PR soon.