From 5c399b64062f09f359668bfc036c2aeedf1e4bb7 Mon Sep 17 00:00:00 2001
From: decolua <decoluadt@gmail.com>
Date: Fri, 11 Sep 2026 22:03:48 +0700
Subject: [PATCH] feat(codebuddy-intl,ollama): add DeepSeek-V4.1-Flash

codebuddy-intl: deepseek-v4-flash replaced by deepseek-v4.1-flash (same
gateway catalog as CN) and a capability override so the model keeps the
openai-style reasoning_effort path instead of the vendor-native "deepseek"
thinking shape the gateway rejects. Thinking levels low/high/xhigh.

ollama: add deepseek-v4.1-flash:cloud (verified on ollama.com/api/tags) with
vision + 1M context caps.

Co-Authored-By: Claude Code <noreply@anthropic.com>
---
 open-sse/providers/capabilities.js            | 17 +++++++++++++++++
 open-sse/providers/registry/codebuddy-intl.js |  4 +++-
 open-sse/providers/registry/ollama.js         |  1 +
 open-sse/providers/thinkingLevels.js          |  2 ++
 4 files changed, 23 insertions(+), 1 deletion(-)

diff --git a/open-sse/providers/capabilities.js b/open-sse/providers/capabilities.js
index e011b788..8ad55314 100644
--- a/open-sse/providers/capabilities.js
+++ b/open-sse/providers/capabilities.js
@@ -214,6 +214,13 @@ export const PROVIDER_CAPABILITIES = {
     // contract). maxOutput 128000 per the server's product-config payload.
     "deepseek-v4.1-flash": { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: true, contextWindow: 1000000, maxOutput: 128000 },
   },
+  // CodeBuddy intl — same gateway catalog as CN, so deepseek-v4.1-flash mirrors
+  // the codebuddy-cn entry (the openai-style reasoning_effort format matters:
+  // the generic *deepseek-v4* pattern would otherwise pick the vendor-native
+  // "deepseek" thinking shape, which the CodeBuddy gateway does not accept).
+  "codebuddy-intl": {
+    "deepseek-v4.1-flash": { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: true, contextWindow: 1000000, maxOutput: 128000 },
+  },
   // Qoder — upstream exposes opaque internal ids (dfmodel, kmodel, …); the
   // registry `name` is display-only and capability lookup matches on the raw
   // id, so every qoder model would fall through to DEFAULT_CAPABILITIES
@@ -251,6 +258,16 @@ export const PROVIDER_CAPABILITIES = {
     "laguna-s-2.1":  { reasoning: true, thinkingFormat: "openai", contextWindow: 1000000, maxOutput: 32000 },
     "laguna-xs-2.1": { reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 32000 },
   },
+  // Ollama Cloud — the generic *deepseek-v4* pattern misses the vision badge
+  // the library page publishes for this model (text+image in, 1M context).
+  // ponytail: thinkingFormat stays "deepseek" to preserve today's body shape;
+  // Ollama's native toggle is the top-level `think` field (bool or
+  // low/medium/high/max), which no format in thinkingUnified.js emits yet —
+  // openai-to-ollama.js drops it. Wire a "think" format when thinking on
+  // Ollama Cloud is actually needed.
+  "ollama": {
+    "deepseek-v4.1-flash:cloud": { vision: true, reasoning: true, thinkingFormat: "deepseek", contextWindow: 1000000, maxOutput: 384000 },
+  },
 };
 
 /**
diff --git a/open-sse/providers/registry/codebuddy-intl.js b/open-sse/providers/registry/codebuddy-intl.js
index eab1ce93..65e787d1 100644
--- a/open-sse/providers/registry/codebuddy-intl.js
+++ b/open-sse/providers/registry/codebuddy-intl.js
@@ -58,7 +58,9 @@ export default {
     { id: "kimi-k2.5", name: "Kimi-K2.5" },
     { id: "hy3-preview", name: "Hy3 Preview" },
     { id: "deepseek-v4-pro", name: "DeepSeek-V4-Pro" },
-    { id: "deepseek-v4-flash", name: "DeepSeek-V4-Flash" },
+    // deepseek-v4-flash replaced server-side by deepseek-v4.1-flash (same
+    // catalog as CN; the old endpoint still answers 200 but the list is the contract).
+    { id: "deepseek-v4.1-flash", name: "DeepSeek-V4.1-Flash" },
     { id: "deepseek-v3-2-volc", name: "DeepSeek-V3.2" },
   ],
   oauth: {
diff --git a/open-sse/providers/registry/ollama.js b/open-sse/providers/registry/ollama.js
index 6965484a..605c4b89 100644
--- a/open-sse/providers/registry/ollama.js
+++ b/open-sse/providers/registry/ollama.js
@@ -30,6 +30,7 @@ export default {
     { id: "glm-4.7-flash", name: "GLM 4.7 Flash" },
     { id: "qwen3.5", name: "Qwen3.5" },
     { id: "minimax-m3", name: "MiniMax M3" },
+    { id: "deepseek-v4.1-flash:cloud", name: "DeepSeek V4.1 Flash" },
   ],
   serviceKinds: ["llm", "webFetch"],
   fetchConfig: {
diff --git a/open-sse/providers/thinkingLevels.js b/open-sse/providers/thinkingLevels.js
index 7cc1b28f..404fa75e 100644
--- a/open-sse/providers/thinkingLevels.js
+++ b/open-sse/providers/thinkingLevels.js
@@ -52,6 +52,8 @@ const PATTERN_THINKING = [
   { 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"] },
+  // codebuddy-intl rides the same gateway catalog, so its deepseek levels match.
+  { provider: "codebuddy-intl", pattern: "deepseek-v4*", levels: ["low", "high", "xhigh"] },
 ];
 
 // Returns valid thinking levels for a model, or null when the model has no reasoning.