feat(codebuddy-cn): refresh model catalog — add hy4-preview/hy3/glm-5.3/kimi-k3, drop EOL glm-5.0/glm-4.7

- Add hy3, hy3-x, hy4-preview, hy4-preview-x, glm-5.3, glm-5.3-flash, kimi-k3-1
- Remove dead models glm-5.0, glm-4.7 (API 11102)
- Register capabilities and context windows in PROVIDER_CAPABILITIES
- Configure supported effort sets in PATTERN_THINKING

open-sse/providers/thinkingLevels.js

@@ -39,6 +39,13 @@ const PATTERN_THINKING = [
   { provider: "codex", pattern: "*gpt-5.6-terra*", levels: [...CODEX_GPT_5_6_LEVELS, "ultra"] },
   { provider: "codex", pattern: "*gpt-5.6-luna*", levels: CODEX_GPT_5_6_LEVELS },
   { pattern: "*codex*", levels: ["low", "medium", "high", "xhigh"] }, // codex cannot disable thinking
+  // codebuddy-cn per-model effort sets — read off the client picker (server-
+  // delivered supportedEfforts), 2026-08-30. Gateway uses thinkingFormat "openai"
+  // but rejects levels outside each model's set.
+  { provider: "codebuddy-cn", pattern: "glm-5.3*",     levels: ["low", "high", "max"] },
+  { provider: "codebuddy-cn", pattern: "deepseek-v4*", levels: ["low", "high", "xhigh"] },
+  { provider: "codebuddy-cn", pattern: "hy3*",         levels: ["low", "high"] },
+  { provider: "codebuddy-cn", pattern: "hy4*",         levels: ["high"] },
 ];
 
 // Returns valid thinking levels for a model, or null when the model has no reasoning.