fix(translator): separate tool calls from content in OpenAI Claude requests

Added detailed operational instructions for Codex agents based on GPT-5, covering shell usage, editing constraints, sandboxing policies, and approval mechanisms. Also included comprehensive review process guidelines for flagging and communicating issues effectively.

.github/workflows/pr-path-guard.yml

@@ -0,0 +1,28 @@
+name: translator-path-guard
+
+on:
+  pull_request:
+    types:
+      - opened
+      - synchronize
+      - reopened
+
+jobs:
+  ensure-no-translator-changes:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - name: Detect internal/translator changes
+        id: changed-files
+        uses: tj-actions/changed-files@v45
+        with:
+          files: |
+            internal/translator/**
+      - name: Fail when restricted paths change
+        if: steps.changed-files.outputs.any_changed == 'true'
+        run: |
+          echo "Changes under internal/translator are not allowed in pull requests."
+          echo "You need to create an issue for our maintenance team to make the necessary changes."
+          exit 1

README.md

@@ -8,6 +8,16 @@ It now also supports OpenAI Codex (GPT models) and Claude Code via OAuth.
 
 So you can use local or multi-account CLI access with OpenAI(include Responses)/Gemini/Claude-compatible clients and SDKs.
 
+## Sponsor
+
+[![z.ai](https://assets.router-for.me/english.png)](https://z.ai/subscribe?ic=8JVLJQFSKB)
+
+This project is sponsored by Z.ai, supporting us with their GLM CODING PLAN.
+
+GLM CODING PLAN is a subscription service designed for AI coding, starting at just $3/month. It provides access to their flagship GLM-4.6 model across 10+ popular AI coding tools (Claude Code, Cline, Roo Code, etc.), offering developers top-tier, fast, and stable coding experiences.
+
+Get 10% OFF GLM CODING PLAN：https://z.ai/subscribe?ic=8JVLJQFSKB
+
 ## Overview
 
 - OpenAI/Gemini/Claude compatible API endpoints for CLI models

README_CN.md

@@ -8,7 +8,15 @@
 
 您可以使用本地或多账户的CLI方式，通过任何与 OpenAI（包括Responses）/Gemini/Claude 兼容的客户端和SDK进行访问。
 
-现已新增国内提供商：[Qwen Code](https://github.com/QwenLM/qwen-code)、[iFlow](https://iflow.cn/)。
+## 赞助商
+
+[![bigmodel.cn](https://assets.router-for.me/chinese.png)](https://www.bigmodel.cn/claude-code?ic=RRVJPB5SII)
+
+本项目由 Z智谱 提供赞助, 他们通过 GLM CODING PLAN 对本项目提供技术支持。
+
+GLM CODING PLAN 是专为AI编码打造的订阅套餐，每月最低仅需20元，即可在十余款主流AI编码工具如 Claude Code、Cline、Roo Code 中畅享智谱旗舰模型GLM-4.6，为开发者提供顶尖的编码体验。
+
+智谱AI为本软件提供了特别优惠，使用以下链接购买可以享受九折优惠：https://www.bigmodel.cn/claude-code?ic=RRVJPB5SII
 
 ## 功能特性
 

internal/api/handlers/management/auth_files.go

@@ -12,6 +12,7 @@ import (
 	"net/url"
 	"os"
 	"path/filepath"
+	"sort"
 	"strconv"
 	"strings"
 	"sync"
@@ -229,8 +230,32 @@ func (h *Handler) managementCallbackURL(path string) (string, error) {
 	return fmt.Sprintf("http://127.0.0.1:%d%s", h.cfg.Port, path), nil
 }
 
-// List auth files
 func (h *Handler) ListAuthFiles(c *gin.Context) {
+	if h == nil {
+		c.JSON(500, gin.H{"error": "handler not initialized"})
+		return
+	}
+	if h.authManager == nil {
+		h.listAuthFilesFromDisk(c)
+		return
+	}
+	auths := h.authManager.List()
+	files := make([]gin.H, 0, len(auths))
+	for _, auth := range auths {
+		if entry := h.buildAuthFileEntry(auth); entry != nil {
+			files = append(files, entry)
+		}
+	}
+	sort.Slice(files, func(i, j int) bool {
+		nameI, _ := files[i]["name"].(string)
+		nameJ, _ := files[j]["name"].(string)
+		return strings.ToLower(nameI) < strings.ToLower(nameJ)
+	})
+	c.JSON(200, gin.H{"files": files})
+}
+
+// List auth files from disk when the auth manager is unavailable.
+func (h *Handler) listAuthFilesFromDisk(c *gin.Context) {
 	entries, err := os.ReadDir(h.cfg.AuthDir)
 	if err != nil {
 		c.JSON(500, gin.H{"error": fmt.Sprintf("failed to read auth dir: %v", err)})
@@ -263,6 +288,103 @@ func (h *Handler) ListAuthFiles(c *gin.Context) {
 	c.JSON(200, gin.H{"files": files})
 }
 
+func (h *Handler) buildAuthFileEntry(auth *coreauth.Auth) gin.H {
+	if auth == nil {
+		return nil
+	}
+	runtimeOnly := isRuntimeOnlyAuth(auth)
+	path := strings.TrimSpace(authAttribute(auth, "path"))
+	if path == "" && !runtimeOnly {
+		return nil
+	}
+	name := strings.TrimSpace(auth.FileName)
+	if name == "" {
+		name = auth.ID
+	}
+	entry := gin.H{
+		"id":             auth.ID,
+		"name":           name,
+		"type":           strings.TrimSpace(auth.Provider),
+		"provider":       strings.TrimSpace(auth.Provider),
+		"label":          auth.Label,
+		"status":         auth.Status,
+		"status_message": auth.StatusMessage,
+		"disabled":       auth.Disabled,
+		"unavailable":    auth.Unavailable,
+		"runtime_only":   runtimeOnly,
+		"source":         "memory",
+		"size":           int64(0),
+	}
+	if email := authEmail(auth); email != "" {
+		entry["email"] = email
+	}
+	if accountType, account := auth.AccountInfo(); accountType != "" || account != "" {
+		if accountType != "" {
+			entry["account_type"] = accountType
+		}
+		if account != "" {
+			entry["account"] = account
+		}
+	}
+	if !auth.CreatedAt.IsZero() {
+		entry["created_at"] = auth.CreatedAt
+	}
+	if !auth.UpdatedAt.IsZero() {
+		entry["modtime"] = auth.UpdatedAt
+		entry["updated_at"] = auth.UpdatedAt
+	}
+	if !auth.LastRefreshedAt.IsZero() {
+		entry["last_refresh"] = auth.LastRefreshedAt
+	}
+	if path != "" {
+		entry["path"] = path
+		entry["source"] = "file"
+		if info, err := os.Stat(path); err == nil {
+			entry["size"] = info.Size()
+			entry["modtime"] = info.ModTime()
+		} else if os.IsNotExist(err) {
+			entry["source"] = "memory"
+		} else {
+			log.WithError(err).Warnf("failed to stat auth file %s", path)
+		}
+	}
+	return entry
+}
+
+func authEmail(auth *coreauth.Auth) string {
+	if auth == nil {
+		return ""
+	}
+	if auth.Metadata != nil {
+		if v, ok := auth.Metadata["email"].(string); ok {
+			return strings.TrimSpace(v)
+		}
+	}
+	if auth.Attributes != nil {
+		if v := strings.TrimSpace(auth.Attributes["email"]); v != "" {
+			return v
+		}
+		if v := strings.TrimSpace(auth.Attributes["account_email"]); v != "" {
+			return v
+		}
+	}
+	return ""
+}
+
+func authAttribute(auth *coreauth.Auth, key string) string {
+	if auth == nil || len(auth.Attributes) == 0 {
+		return ""
+	}
+	return auth.Attributes[key]
+}
+
+func isRuntimeOnlyAuth(auth *coreauth.Auth) bool {
+	if auth == nil || len(auth.Attributes) == 0 {
+		return false
+	}
+	return strings.EqualFold(strings.TrimSpace(auth.Attributes["runtime_only"]), "true")
+}
+
 // Download single auth file by name
 func (h *Handler) DownloadAuthFile(c *gin.Context) {
 	name := c.Query("name")

internal/misc/codex_instructions/gpt_5_codex_prompt.md-007-8c75ed39d5bb94159d21072d7384765d94a9012b

@@ -0,0 +1,107 @@
+You are Codex, based on GPT-5. You are running as a coding agent in the Codex CLI on a user's computer.
+
+## General
+
+- The arguments to `shell` will be passed to execvp(). Most terminal commands should be prefixed with ["bash", "-lc"].
+- Always set the `workdir` param when using the shell function. Do not use `cd` unless absolutely necessary.
+- When searching for text or files, prefer using `rg` or `rg --files` respectively because `rg` is much faster than alternatives like `grep`. (If the `rg` command is not found, then use alternatives.)
+
+## Editing constraints
+
+- Default to ASCII when editing or creating files. Only introduce non-ASCII or other Unicode characters when there is a clear justification and the file already uses them.
+- Add succinct code comments that explain what is going on if code is not self-explanatory. You should not add comments like "Assigns the value to the variable", but a brief comment might be useful ahead of a complex code block that the user would otherwise have to spend time parsing out. Usage of these comments should be rare.
+- Try to use apply_patch for single file edits, but it is fine to explore other options to make the edit if it does not work well. Do not use apply_patch for changes that are auto-generated (i.e. generating package.json or running a lint or format command like gofmt) or when scripting is more efficient (such as search and replacing a string across a codebase).
+- You may be in a dirty git worktree.
+    * NEVER revert existing changes you did not make unless explicitly requested, since these changes were made by the user.
+    * If asked to make a commit or code edits and there are unrelated changes to your work or changes that you didn't make in those files, don't revert those changes.
+    * If the changes are in files you've touched recently, you should read carefully and understand how you can work with the changes rather than reverting them.
+    * If the changes are in unrelated files, just ignore them and don't revert them.
+- Do not amend a commit unless explicitly requested to do so.
+- While you are working, you might notice unexpected changes that you didn't make. If this happens, STOP IMMEDIATELY and ask the user how they would like to proceed.
+- **NEVER** use destructive commands like `git reset --hard` or `git checkout --` unless specifically requested or approved by the user.
+
+## Plan tool
+
+When using the planning tool:
+- Skip using the planning tool for straightforward tasks (roughly the easiest 25%).
+- Do not make single-step plans.
+- When you made a plan, update it after having performed one of the sub-tasks that you shared on the plan.
+
+## Codex CLI harness, sandboxing, and approvals
+
+The Codex CLI harness supports several different configurations for sandboxing and escalation approvals that the user can choose from.
+
+Filesystem sandboxing defines which files can be read or written. The options for `sandbox_mode` are:
+- **read-only**: The sandbox only permits reading files.
+- **workspace-write**: The sandbox permits reading files, and editing files in `cwd` and `writable_roots`. Editing files in other directories requires approval.
+- **danger-full-access**: No filesystem sandboxing - all commands are permitted.
+
+Network sandboxing defines whether network can be accessed without approval. Options for `network_access` are:
+- **restricted**: Requires approval
+- **enabled**: No approval needed
+
+Approvals are your mechanism to get user consent to run shell commands without the sandbox. Possible configuration options for `approval_policy` are
+- **untrusted**: The harness will escalate most commands for user approval, apart from a limited allowlist of safe "read" commands.
+- **on-failure**: The harness will allow all commands to run in the sandbox (if enabled), and failures will be escalated to the user for approval to run again without the sandbox.
+- **on-request**: Commands will be run in the sandbox by default, and you can specify in your tool call if you want to escalate a command to run without sandboxing. (Note that this mode is not always available. If it is, you'll see parameters for it in the `shell` command description.)
+- **never**: This is a non-interactive mode where you may NEVER ask the user for approval to run commands. Instead, you must always persist and work around constraints to solve the task for the user. You MUST do your utmost best to finish the task and validate your work before yielding. If this mode is paired with `danger-full-access`, take advantage of it to deliver the best outcome for the user. Further, in this mode, your default testing philosophy is overridden: Even if you don't see local patterns for testing, you may add tests and scripts to validate your work. Just remove them before yielding.
+
+When you are running with `approval_policy == on-request`, and sandboxing enabled, here are scenarios where you'll need to request approval:
+- You need to run a command that writes to a directory that requires it (e.g. running tests that write to /var)
+- You need to run a GUI app (e.g., open/xdg-open/osascript) to open browsers or files.
+- You are running sandboxed and need to run a command that requires network access (e.g. installing packages)
+- If you run a command that is important to solving the user's query, but it fails because of sandboxing, rerun the command with approval. ALWAYS proceed to use the `with_escalated_permissions` and `justification` parameters - do not message the user before requesting approval for the command.
+- You are about to take a potentially destructive action such as an `rm` or `git reset` that the user did not explicitly ask for
+- (for all of these, you should weigh alternative paths that do not require approval)
+
+When `sandbox_mode` is set to read-only, you'll need to request approval for any command that isn't a read.
+
+You will be told what filesystem sandboxing, network sandboxing, and approval mode are active in a developer or user message. If you are not told about this, assume that you are running with workspace-write, network sandboxing enabled, and approval on-failure.
+
+Although they introduce friction to the user because your work is paused until the user responds, you should leverage them when necessary to accomplish important work. If the completing the task requires escalated permissions, Do not let these settings or the sandbox deter you from attempting to accomplish the user's task unless it is set to "never", in which case never ask for approvals.
+
+When requesting approval to execute a command that will require escalated privileges:
+  - Provide the `with_escalated_permissions` parameter with the boolean value true
+  - Include a short, 1 sentence explanation for why you need to enable `with_escalated_permissions` in the justification parameter
+
+## Special user requests
+
+- If the user makes a simple request (such as asking for the time) which you can fulfill by running a terminal command (such as `date`), you should do so.
+- If the user asks for a "review", default to a code review mindset: prioritise identifying bugs, risks, behavioural regressions, and missing tests. Findings must be the primary focus of the response - keep summaries or overviews brief and only after enumerating the issues. Present findings first (ordered by severity with file/line references), follow with open questions or assumptions, and offer a change-summary only as a secondary detail. If no findings are discovered, state that explicitly and mention any residual risks or testing gaps.
+
+## Presenting your work and final message
+
+You are producing plain text that will later be styled by the CLI. Follow these rules exactly. Formatting should make results easy to scan, but not feel mechanical. Use judgment to decide how much structure adds value.
+
+- Default: be very concise; friendly coding teammate tone.
+- Ask only when needed; suggest ideas; mirror the user's style.
+- For substantial work, summarize clearly; follow final‑answer formatting.
+- Skip heavy formatting for simple confirmations.
+- Don't dump large files you've written; reference paths only.
+- No "save/copy this file" - User is on the same machine.
+- Offer logical next steps (tests, commits, build) briefly; add verify steps if you couldn't do something.
+- For code changes:
+  * Lead with a quick explanation of the change, and then give more details on the context covering where and why a change was made. Do not start this explanation with "summary", just jump right in.
+  * If there are natural next steps the user may want to take, suggest them at the end of your response. Do not make suggestions if there are no natural next steps.
+  * When suggesting multiple options, use numeric lists for the suggestions so the user can quickly respond with a single number.
+- The user does not command execution outputs. When asked to show the output of a command (e.g. `git show`), relay the important details in your answer or summarize the key lines so the user understands the result.
+
+### Final answer structure and style guidelines
+
+- Plain text; CLI handles styling. Use structure only when it helps scanability.
+- Headers: optional; short Title Case (1-3 words) wrapped in **…**; no blank line before the first bullet; add only if they truly help.
+- Bullets: use - ; merge related points; keep to one line when possible; 4–6 per list ordered by importance; keep phrasing consistent.
+- Monospace: backticks for commands/paths/env vars/code ids and inline examples; use for literal keyword bullets; never combine with **.
+- Code samples or multi-line snippets should be wrapped in fenced code blocks; include an info string as often as possible.
+- Structure: group related bullets; order sections general → specific → supporting; for subsections, start with a bolded keyword bullet, then items; match complexity to the task.
+- Tone: collaborative, concise, factual; present tense, active voice; self‑contained; no "above/below"; parallel wording.
+- Don'ts: no nested bullets/hierarchies; no ANSI codes; don't cram unrelated keywords; keep keyword lists short—wrap/reformat if long; avoid naming formatting styles in answers.
+- Adaptation: code explanations → precise, structured with code refs; simple tasks → lead with outcome; big changes → logical walkthrough + rationale + next actions; casual one-offs → plain sentences, no headers/bullets.
+- File References: When referencing files in your response, make sure to include the relevant start line and always follow the below rules:
+  * Use inline code to make file paths clickable.
+  * Each reference should have a stand alone path. Even if it's the same file.
+  * Accepted: absolute, workspace‑relative, a/ or b/ diff prefixes, or bare filename/suffix.
+  * Line/column (1‑based, optional): :line[:column] or #Lline[Ccolumn] (column defaults to 1).
+  * Do not use URIs like file://, vscode://, or https://.
+  * Do not provide range of lines
+  * Examples: src/app.ts, src/app.ts:42, b/server/index.js#L10, C:\repo\project\main.rs:12:5

internal/misc/codex_instructions/review_prompt.md-002-f842849bec97326ad6fb40e9955b6ba9f0f3fc0d

@@ -0,0 +1,87 @@
+# Review guidelines:
+
+You are acting as a reviewer for a proposed code change made by another engineer.
+
+Below are some default guidelines for determining whether the original author would appreciate the issue being flagged.
+
+These are not the final word in determining whether an issue is a bug. In many cases, you will encounter other, more specific guidelines. These may be present elsewhere in a developer message, a user message, a file, or even elsewhere in this system message.
+Those guidelines should be considered to override these general instructions.
+
+Here are the general guidelines for determining whether something is a bug and should be flagged.
+
+1. It meaningfully impacts the accuracy, performance, security, or maintainability of the code.
+2. The bug is discrete and actionable (i.e. not a general issue with the codebase or a combination of multiple issues).
+3. Fixing the bug does not demand a level of rigor that is not present in the rest of the codebase (e.g. one doesn't need very detailed comments and input validation in a repository of one-off scripts in personal projects)
+4. The bug was introduced in the commit (pre-existing bugs should not be flagged).
+5. The author of the original PR would likely fix the issue if they were made aware of it.
+6. The bug does not rely on unstated assumptions about the codebase or author's intent.
+7. It is not enough to speculate that a change may disrupt another part of the codebase, to be considered a bug, one must identify the other parts of the code that are provably affected.
+8. The bug is clearly not just an intentional change by the original author.
+
+When flagging a bug, you will also provide an accompanying comment. Once again, these guidelines are not the final word on how to construct a comment -- defer to any subsequent guidelines that you encounter.
+
+1. The comment should be clear about why the issue is a bug.
+2. The comment should appropriately communicate the severity of the issue. It should not claim that an issue is more severe than it actually is.
+3. The comment should be brief. The body should be at most 1 paragraph. It should not introduce line breaks within the natural language flow unless it is necessary for the code fragment.
+4. The comment should not include any chunks of code longer than 3 lines. Any code chunks should be wrapped in markdown inline code tags or a code block.
+5. The comment should clearly and explicitly communicate the scenarios, environments, or inputs that are necessary for the bug to arise. The comment should immediately indicate that the issue's severity depends on these factors.
+6. The comment's tone should be matter-of-fact and not accusatory or overly positive. It should read as a helpful AI assistant suggestion without sounding too much like a human reviewer.
+7. The comment should be written such that the original author can immediately grasp the idea without close reading.
+8. The comment should avoid excessive flattery and comments that are not helpful to the original author. The comment should avoid phrasing like "Great job ...", "Thanks for ...".
+
+Below are some more detailed guidelines that you should apply to this specific review.
+
+HOW MANY FINDINGS TO RETURN:
+
+Output all findings that the original author would fix if they knew about it. If there is no finding that a person would definitely love to see and fix, prefer outputting no findings. Do not stop at the first qualifying finding. Continue until you've listed every qualifying finding.
+
+GUIDELINES:
+
+- Ignore trivial style unless it obscures meaning or violates documented standards.
+- Use one comment per distinct issue (or a multi-line range if necessary).
+- Use ```suggestion blocks ONLY for concrete replacement code (minimal lines; no commentary inside the block).
+- In every ```suggestion block, preserve the exact leading whitespace of the replaced lines (spaces vs tabs, number of spaces).
+- Do NOT introduce or remove outer indentation levels unless that is the actual fix.
+
+The comments will be presented in the code review as inline comments. You should avoid providing unnecessary location details in the comment body. Always keep the line range as short as possible for interpreting the issue. Avoid ranges longer than 5–10 lines; instead, choose the most suitable subrange that pinpoints the problem.
+
+At the beginning of the finding title, tag the bug with priority level. For example "[P1] Un-padding slices along wrong tensor dimensions". [P0] – Drop everything to fix.  Blocking release, operations, or major usage. Only use for universal issues that do not depend on any assumptions about the inputs. · [P1] – Urgent. Should be addressed in the next cycle · [P2] – Normal. To be fixed eventually · [P3] – Low. Nice to have.
+
+Additionally, include a numeric priority field in the JSON output for each finding: set "priority" to 0 for P0, 1 for P1, 2 for P2, or 3 for P3. If a priority cannot be determined, omit the field or use null.
+
+At the end of your findings, output an "overall correctness" verdict of whether or not the patch should be considered "correct".
+Correct implies that existing code and tests will not break, and the patch is free of bugs and other blocking issues.
+Ignore non-blocking issues such as style, formatting, typos, documentation, and other nits.
+
+FORMATTING GUIDELINES:
+The finding description should be one paragraph.
+
+OUTPUT FORMAT:
+
+## Output schema  — MUST MATCH *exactly*
+
+```json
+{
+  "findings": [
+    {
+      "title": "<≤ 80 chars, imperative>",
+      "body": "<valid Markdown explaining *why* this is a problem; cite files/lines/functions>",
+      "confidence_score": <float 0.0-1.0>,
+      "priority": <int 0-3, optional>,
+      "code_location": {
+        "absolute_file_path": "<file path>",
+        "line_range": {"start": <int>, "end": <int>}
+      }
+    }
+  ],
+  "overall_correctness": "patch is correct" | "patch is incorrect",
+  "overall_explanation": "<1-3 sentence explanation justifying the overall_correctness verdict>",
+  "overall_confidence_score": <float 0.0-1.0>
+}
+```
+
+* **Do not** wrap the JSON in markdown fences or extra prose.
+* The code_location field is required and must include absolute_file_path and line_range.
+* Line ranges must be as short as possible for interpreting the issue (avoid ranges over 5–10 lines; pick the most suitable subrange).
+* The code_location should overlap with the diff.
+* Do not generate a PR fix.

internal/registry/model_definitions.go

@@ -116,36 +116,6 @@ func GeminiModels() []*ModelInfo {
 			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
 			Thinking:                   &ThinkingSupport{Min: 512, Max: 24576, ZeroAllowed: true, DynamicAllowed: true},
 		},
-		{
-			ID:                         "gemini-2.5-flash-image-preview",
-			Object:                     "model",
-			Created:                    time.Now().Unix(),
-			OwnedBy:                    "google",
-			Type:                       "gemini",
-			Name:                       "models/gemini-2.5-flash-image-preview",
-			Version:                    "2.5",
-			DisplayName:                "Gemini 2.5 Flash Image Preview",
-			Description:                "State-of-the-art image generation and editing model.",
-			InputTokenLimit:            1048576,
-			OutputTokenLimit:           8192,
-			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
-			// image models don't support thinkingConfig; leave Thinking nil
-		},
-		{
-			ID:                         "gemini-2.5-flash-image",
-			Object:                     "model",
-			Created:                    time.Now().Unix(),
-			OwnedBy:                    "google",
-			Type:                       "gemini",
-			Name:                       "models/gemini-2.5-flash-image",
-			Version:                    "2.5",
-			DisplayName:                "Gemini 2.5 Flash Image",
-			Description:                "State-of-the-art image generation and editing model.",
-			InputTokenLimit:            1048576,
-			OutputTokenLimit:           8192,
-			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
-			// image models don't support thinkingConfig; leave Thinking nil
-		},
 	}
 }
 
@@ -153,57 +123,138 @@ func GeminiModels() []*ModelInfo {
 func GetGeminiModels() []*ModelInfo { return GeminiModels() }
 
 // GetGeminiCLIModels returns the standard Gemini model definitions
-func GetGeminiCLIModels() []*ModelInfo { return GeminiModels() }
-
-// GetAIStudioModels returns the Gemini model definitions for AI Studio integrations
-func GetAIStudioModels() []*ModelInfo {
-	base := GeminiModels()
-	return append(base,
-		&ModelInfo{
-			ID:                         "gemini-pro-latest",
+func GetGeminiCLIModels() []*ModelInfo {
+	return []*ModelInfo{
+		{
+			ID:                         "gemini-2.5-flash",
 			Object:                     "model",
 			Created:                    time.Now().Unix(),
 			OwnedBy:                    "google",
 			Type:                       "gemini",
-			Name:                       "models/gemini-pro-latest",
-			Version:                    "2.5",
-			DisplayName:                "Gemini Pro Latest",
-			Description:                "Latest release of Gemini Pro",
-			InputTokenLimit:            1048576,
-			OutputTokenLimit:           65536,
-			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
-			Thinking:                   &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true},
-		},
-		&ModelInfo{
-			ID:                         "gemini-flash-latest",
-			Object:                     "model",
-			Created:                    time.Now().Unix(),
-			OwnedBy:                    "google",
-			Type:                       "gemini",
-			Name:                       "models/gemini-flash-latest",
-			Version:                    "2.5",
-			DisplayName:                "Gemini Flash Latest",
-			Description:                "Latest release of Gemini Flash",
+			Name:                       "models/gemini-2.5-flash",
+			Version:                    "001",
+			DisplayName:                "Gemini 2.5 Flash",
+			Description:                "Stable version of Gemini 2.5 Flash, our mid-size multimodal model that supports up to 1 million tokens, released in June of 2025.",
 			InputTokenLimit:            1048576,
 			OutputTokenLimit:           65536,
 			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
 			Thinking:                   &ThinkingSupport{Min: 0, Max: 24576, ZeroAllowed: true, DynamicAllowed: true},
 		},
-		&ModelInfo{
-			ID:                         "gemini-flash-lite-latest",
+		{
+			ID:                         "gemini-2.5-pro",
 			Object:                     "model",
 			Created:                    time.Now().Unix(),
 			OwnedBy:                    "google",
 			Type:                       "gemini",
-			Name:                       "models/gemini-flash-lite-latest",
+			Name:                       "models/gemini-2.5-pro",
 			Version:                    "2.5",
-			DisplayName:                "Gemini Flash-Lite Latest",
-			Description:                "Latest release of Gemini Flash-Lite",
+			DisplayName:                "Gemini 2.5 Pro",
+			Description:                "Stable release (June 17th, 2025) of Gemini 2.5 Pro",
 			InputTokenLimit:            1048576,
 			OutputTokenLimit:           65536,
 			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
-			Thinking:                   &ThinkingSupport{Min: 512, Max: 24576, ZeroAllowed: true, DynamicAllowed: true},
+			Thinking:                   &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true},
 		},
+		{
+			ID:                         "gemini-3-pro-preview-11-2025",
+			Object:                     "model",
+			Created:                    time.Now().Unix(),
+			OwnedBy:                    "google",
+			Type:                       "gemini",
+			Name:                       "models/gemini-3-pro-preview-11-2025",
+			Version:                    "3",
+			DisplayName:                "Gemini 3 Pro Preview 11-2025",
+			Description:                "Latest preview of Gemini Pro",
+			InputTokenLimit:            1048576,
+			OutputTokenLimit:           65536,
+			SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+			Thinking:                   &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true},
+		},
+	}
+}
+
+// GetAIStudioModels returns the Gemini model definitions for AI Studio integrations
+func GetAIStudioModels() []*ModelInfo {
+	base := GeminiModels()
+
+	return append(base,
+		[]*ModelInfo{
+			{
+				ID:                         "gemini-pro-latest",
+				Object:                     "model",
+				Created:                    time.Now().Unix(),
+				OwnedBy:                    "google",
+				Type:                       "gemini",
+				Name:                       "models/gemini-pro-latest",
+				Version:                    "2.5",
+				DisplayName:                "Gemini Pro Latest",
+				Description:                "Latest release of Gemini Pro",
+				InputTokenLimit:            1048576,
+				OutputTokenLimit:           65536,
+				SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+				Thinking:                   &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true},
+			},
+			{
+				ID:                         "gemini-flash-latest",
+				Object:                     "model",
+				Created:                    time.Now().Unix(),
+				OwnedBy:                    "google",
+				Type:                       "gemini",
+				Name:                       "models/gemini-flash-latest",
+				Version:                    "2.5",
+				DisplayName:                "Gemini Flash Latest",
+				Description:                "Latest release of Gemini Flash",
+				InputTokenLimit:            1048576,
+				OutputTokenLimit:           65536,
+				SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+				Thinking:                   &ThinkingSupport{Min: 0, Max: 24576, ZeroAllowed: true, DynamicAllowed: true},
+			},
+			{
+				ID:                         "gemini-flash-lite-latest",
+				Object:                     "model",
+				Created:                    time.Now().Unix(),
+				OwnedBy:                    "google",
+				Type:                       "gemini",
+				Name:                       "models/gemini-flash-lite-latest",
+				Version:                    "2.5",
+				DisplayName:                "Gemini Flash-Lite Latest",
+				Description:                "Latest release of Gemini Flash-Lite",
+				InputTokenLimit:            1048576,
+				OutputTokenLimit:           65536,
+				SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+				Thinking:                   &ThinkingSupport{Min: 512, Max: 24576, ZeroAllowed: true, DynamicAllowed: true},
+			},
+			{
+				ID:                         "gemini-2.5-flash-image-preview",
+				Object:                     "model",
+				Created:                    time.Now().Unix(),
+				OwnedBy:                    "google",
+				Type:                       "gemini",
+				Name:                       "models/gemini-2.5-flash-image-preview",
+				Version:                    "2.5",
+				DisplayName:                "Gemini 2.5 Flash Image Preview",
+				Description:                "State-of-the-art image generation and editing model.",
+				InputTokenLimit:            1048576,
+				OutputTokenLimit:           8192,
+				SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+				// image models don't support thinkingConfig; leave Thinking nil
+			},
+			{
+				ID:                         "gemini-2.5-flash-image",
+				Object:                     "model",
+				Created:                    time.Now().Unix(),
+				OwnedBy:                    "google",
+				Type:                       "gemini",
+				Name:                       "models/gemini-2.5-flash-image",
+				Version:                    "2.5",
+				DisplayName:                "Gemini 2.5 Flash Image",
+				Description:                "State-of-the-art image generation and editing model.",
+				InputTokenLimit:            1048576,
+				OutputTokenLimit:           8192,
+				SupportedGenerationMethods: []string{"generateContent", "countTokens", "createCachedContent", "batchGenerateContent"},
+				// image models don't support thinkingConfig; leave Thinking nil
+			},
+		}...,
 	)
 }
 
@@ -328,17 +379,43 @@ func GetOpenAIModels() []*ModelInfo {
 			SupportedParameters: []string{"tools"},
 		},
 		{
-			ID:                  "codex-mini-latest",
+			ID:                  "gpt-5-codex-mini",
 			Object:              "model",
 			Created:             time.Now().Unix(),
 			OwnedBy:             "openai",
 			Type:                "openai",
-			Version:             "1.0",
-			DisplayName:         "Codex Mini",
-			Description:         "Lightweight code generation model",
-			ContextLength:       4096,
-			MaxCompletionTokens: 2048,
-			SupportedParameters: []string{"temperature", "max_tokens", "stream", "stop"},
+			Version:             "gpt-5-2025-11-07",
+			DisplayName:         "GPT 5 Codex Mini",
+			Description:         "Stable version of GPT 5 Codex Mini: cheaper, faster, but less capable version of GPT 5 Codex.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5-codex-mini-medium",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5-2025-11-07",
+			DisplayName:         "GPT 5 Codex Mini Medium",
+			Description:         "Stable version of GPT 5 Codex Mini: cheaper, faster, but less capable version of GPT 5 Codex.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
+		},
+		{
+			ID:                  "gpt-5-codex-mini-high",
+			Object:              "model",
+			Created:             time.Now().Unix(),
+			OwnedBy:             "openai",
+			Type:                "openai",
+			Version:             "gpt-5-2025-11-07",
+			DisplayName:         "GPT 5 Codex Mini High",
+			Description:         "Stable version of GPT 5 Codex Mini: cheaper, faster, but less capable version of GPT 5 Codex.",
+			ContextLength:       400000,
+			MaxCompletionTokens: 128000,
+			SupportedParameters: []string{"tools"},
 		},
 	}
 }
@@ -414,6 +491,7 @@ func GetIFlowModels() []*ModelInfo {
 		{ID: "qwen3-235b-a22b-thinking-2507", DisplayName: "Qwen3-235B-A22B-Thinking", Description: "Qwen3 235B A22B Thinking (2507)"},
 		{ID: "qwen3-235b-a22b-instruct", DisplayName: "Qwen3-235B-A22B-Instruct", Description: "Qwen3 235B A22B Instruct"},
 		{ID: "qwen3-235b", DisplayName: "Qwen3-235B-A22B", Description: "Qwen3 235B A22B"},
+		{ID: "minimax-m2", DisplayName: "MiniMax-M2", Description: "MiniMax M2"},
 	}
 	models := make([]*ModelInfo, 0, len(entries))
 	for _, entry := range entries {

internal/runtime/executor/codex_executor.go

@@ -75,6 +75,16 @@ func (e *CodexExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, re
 		case "gpt-5-codex-high":
 			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
 		}
+	} else if util.InArray([]string{"gpt-5-codex-mini", "gpt-5-codex-mini-medium", "gpt-5-codex-mini-high"}, req.Model) {
+		body, _ = sjson.SetBytes(body, "model", "gpt-5-codex-mini")
+		switch req.Model {
+		case "gpt-5-codex-mini-medium":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		case "gpt-5-codex-mini-high":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
+		default:
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		}
 	}
 
 	body, _ = sjson.SetBytes(body, "stream", true)
@@ -188,6 +198,14 @@ func (e *CodexExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 		case "gpt-5-codex-high":
 			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
 		}
+	} else if util.InArray([]string{"gpt-5-codex-mini", "gpt-5-codex-mini-medium", "gpt-5-codex-mini-high"}, req.Model) {
+		body, _ = sjson.SetBytes(body, "model", "gpt-5-codex-mini")
+		switch req.Model {
+		case "gpt-5-codex-mini-medium":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		case "gpt-5-codex-mini-high":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
+		}
 	}
 
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
@@ -312,6 +330,17 @@ func (e *CodexExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Auth
 		default:
 			body, _ = sjson.SetBytes(body, "reasoning.effort", "low")
 		}
+	} else if util.InArray([]string{"gpt-5-codex-mini", "gpt-5-codex-mini-medium", "gpt-5-codex-mini-high"}, req.Model) {
+		modelForCounting = "gpt-5"
+		body, _ = sjson.SetBytes(body, "model", "codex-mini-latest")
+		switch req.Model {
+		case "gpt-5-codex-mini-medium":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		case "gpt-5-codex-mini-high":
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "high")
+		default:
+			body, _ = sjson.SetBytes(body, "reasoning.effort", "medium")
+		}
 	}
 
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
@@ -508,6 +537,11 @@ func (e *CodexExecutor) cacheHelper(ctx context.Context, from sdktranslator.Form
 				codexCacheMap[key] = cache
 			}
 		}
+	} else if from == "openai-response" {
+		promptCacheKey := gjson.GetBytes(req.Payload, "prompt_cache_key")
+		if promptCacheKey.Exists() {
+			cache.ID = promptCacheKey.String()
+		}
 	}
 
 	rawJSON, _ = sjson.SetBytes(rawJSON, "prompt_cache_key", cache.ID)

internal/runtime/executor/gemini_cli_executor.go

@@ -634,11 +634,19 @@ func geminiCLIClientMetadata() string {
 func cliPreviewFallbackOrder(model string) []string {
 	switch model {
 	case "gemini-2.5-pro":
-		return []string{"gemini-2.5-pro-preview-05-06", "gemini-2.5-pro-preview-06-05"}
+		return []string{
+			// "gemini-2.5-pro-preview-05-06",
+			"gemini-2.5-pro-preview-06-05",
+		}
 	case "gemini-2.5-flash":
-		return []string{"gemini-2.5-flash-preview-04-17", "gemini-2.5-flash-preview-05-20"}
+		return []string{
+			// "gemini-2.5-flash-preview-04-17",
+			// "gemini-2.5-flash-preview-05-20",
+		}
 	case "gemini-2.5-flash-lite":
-		return []string{"gemini-2.5-flash-lite-preview-06-17"}
+		return []string{
+			// "gemini-2.5-flash-lite-preview-06-17",
+		}
 	default:
 		return nil
 	}

internal/translator/gemini-cli/claude/gemini-cli_claude_request.go

@@ -99,7 +99,7 @@ func ConvertClaudeRequestToCLI(modelName string, inputRawJSON []byte, _ bool) []
 							if len(toolCallIDs) > 1 {
 								funcName = strings.Join(toolCallIDs[0:len(toolCallIDs)-1], "-")
 							}
-							responseData := contentResult.Get("content").String()
+							responseData := contentResult.Get("content").Raw
 							functionResponse := client.FunctionResponse{Name: funcName, Response: map[string]interface{}{"result": responseData}}
 							clientContent.Parts = append(clientContent.Parts, client.Part{FunctionResponse: &functionResponse})
 						}
@@ -127,6 +127,7 @@ func ConvertClaudeRequestToCLI(modelName string, inputRawJSON []byte, _ bool) []
 				inputSchema := inputSchemaResult.Raw
 				tool, _ := sjson.Delete(toolResult.Raw, "input_schema")
 				tool, _ = sjson.SetRaw(tool, "parametersJsonSchema", inputSchema)
+				tool, _ = sjson.Delete(tool, "strict")
 				var toolDeclaration any
 				if err := json.Unmarshal([]byte(tool), &toolDeclaration); err == nil {
 					tools[0].FunctionDeclarations = append(tools[0].FunctionDeclarations, toolDeclaration)

internal/translator/gemini-cli/openai/chat-completions/gemini-cli_openai_request.go

@@ -65,17 +65,23 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 		if tc := gjson.GetBytes(rawJSON, "extra_body.google.thinking_config"); tc.Exists() && tc.IsObject() {
 			var setBudget bool
 			var normalized int
-			if v := tc.Get("thinking_budget"); v.Exists() {
+
+			if v := tc.Get("thinkingBudget"); v.Exists() {
+				normalized = util.NormalizeThinkingBudget(modelName, int(v.Int()))
+				out, _ = sjson.SetBytes(out, "request.generationConfig.thinkingConfig.thinkingBudget", normalized)
+				setBudget = true
+			} else if v := tc.Get("thinking_budget"); v.Exists() {
 				normalized = util.NormalizeThinkingBudget(modelName, int(v.Int()))
 				out, _ = sjson.SetBytes(out, "request.generationConfig.thinkingConfig.thinkingBudget", normalized)
 				setBudget = true
 			}
-			if v := tc.Get("include_thoughts"); v.Exists() {
+
+			if v := tc.Get("includeThoughts"); v.Exists() {
 				out, _ = sjson.SetBytes(out, "request.generationConfig.thinkingConfig.include_thoughts", v.Bool())
-			} else if setBudget {
-				if normalized != 0 {
-					out, _ = sjson.SetBytes(out, "request.generationConfig.thinkingConfig.include_thoughts", true)
-				}
+			} else if v := tc.Get("include_thoughts"); v.Exists() {
+				out, _ = sjson.SetBytes(out, "request.generationConfig.thinkingConfig.include_thoughts", v.Bool())
+			} else if setBudget && normalized != 0 {
+				out, _ = sjson.SetBytes(out, "request.generationConfig.thinkingConfig.include_thoughts", true)
 			}
 		}
 	}
@@ -149,11 +155,7 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 				toolCallID := m.Get("tool_call_id").String()
 				if toolCallID != "" {
 					c := m.Get("content")
-					if c.Type == gjson.String {
-						toolResponses[toolCallID] = c.String()
-					} else if c.IsObject() && c.Get("type").String() == "text" {
-						toolResponses[toolCallID] = c.Get("text").String()
-					}
+					toolResponses[toolCallID] = c.Raw
 				}
 			}
 		}
@@ -254,7 +256,7 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 								if resp == "" {
 									resp = "{}"
 								}
-								toolNode, _ = sjson.SetRawBytes(toolNode, "parts."+itoa(pp)+".functionResponse.response", []byte(`{"result":`+quoteIfNeeded(resp)+`}`))
+								toolNode, _ = sjson.SetBytes(toolNode, "parts."+itoa(pp)+".functionResponse.response.result", []byte(resp))
 								pp++
 							}
 						}

internal/translator/gemini/claude/gemini_claude_request.go

@@ -92,7 +92,7 @@ func ConvertClaudeRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 							if len(toolCallIDs) > 1 {
 								funcName = strings.Join(toolCallIDs[0:len(toolCallIDs)-1], "-")
 							}
-							responseData := contentResult.Get("content").String()
+							responseData := contentResult.Get("content").Raw
 							functionResponse := client.FunctionResponse{Name: funcName, Response: map[string]interface{}{"result": responseData}}
 							clientContent.Parts = append(clientContent.Parts, client.Part{FunctionResponse: &functionResponse})
 						}
@@ -120,6 +120,7 @@ func ConvertClaudeRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 				inputSchema := inputSchemaResult.Raw
 				tool, _ := sjson.Delete(toolResult.Raw, "input_schema")
 				tool, _ = sjson.SetRaw(tool, "parametersJsonSchema", inputSchema)
+				tool, _ = sjson.Delete(tool, "strict")
 				var toolDeclaration any
 				if err := json.Unmarshal([]byte(tool), &toolDeclaration); err == nil {
 					tools[0].FunctionDeclarations = append(tools[0].FunctionDeclarations, toolDeclaration)

internal/translator/gemini/openai/chat-completions/gemini_openai_request.go

@@ -65,18 +65,23 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 		if tc := gjson.GetBytes(rawJSON, "extra_body.google.thinking_config"); tc.Exists() && tc.IsObject() {
 			var setBudget bool
 			var normalized int
-			if v := tc.Get("thinking_budget"); v.Exists() {
-				// Normalize budget to model range
+
+			if v := tc.Get("thinkingBudget"); v.Exists() {
+				normalized = util.NormalizeThinkingBudget(modelName, int(v.Int()))
+				out, _ = sjson.SetBytes(out, "generationConfig.thinkingConfig.thinkingBudget", normalized)
+				setBudget = true
+			} else if v := tc.Get("thinking_budget"); v.Exists() {
 				normalized = util.NormalizeThinkingBudget(modelName, int(v.Int()))
 				out, _ = sjson.SetBytes(out, "generationConfig.thinkingConfig.thinkingBudget", normalized)
 				setBudget = true
 			}
-			if v := tc.Get("include_thoughts"); v.Exists() {
+
+			if v := tc.Get("includeThoughts"); v.Exists() {
 				out, _ = sjson.SetBytes(out, "generationConfig.thinkingConfig.include_thoughts", v.Bool())
-			} else if setBudget {
-				if normalized != 0 {
-					out, _ = sjson.SetBytes(out, "generationConfig.thinkingConfig.include_thoughts", true)
-				}
+			} else if v := tc.Get("include_thoughts"); v.Exists() {
+				out, _ = sjson.SetBytes(out, "generationConfig.thinkingConfig.include_thoughts", v.Bool())
+			} else if setBudget && normalized != 0 {
+				out, _ = sjson.SetBytes(out, "generationConfig.thinkingConfig.include_thoughts", true)
 			}
 		}
 	}
@@ -150,14 +155,11 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 				toolCallID := m.Get("tool_call_id").String()
 				if toolCallID != "" {
 					c := m.Get("content")
-					if c.Type == gjson.String {
-						toolResponses[toolCallID] = c.String()
-					} else if c.IsObject() && c.Get("type").String() == "text" {
-						toolResponses[toolCallID] = c.Get("text").String()
-					}
+					toolResponses[toolCallID] = c.Raw
 				}
 			}
 		}
+		fmt.Printf("11111")
 
 		for i := 0; i < len(arr); i++ {
 			m := arr[i]
@@ -280,7 +282,7 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 								if resp == "" {
 									resp = "{}"
 								}
-								toolNode, _ = sjson.SetRawBytes(toolNode, "parts."+itoa(pp)+".functionResponse.response", []byte(`{"result":`+quoteIfNeeded(resp)+`}`))
+								toolNode, _ = sjson.SetBytes(toolNode, "parts."+itoa(pp)+".functionResponse.response.result", []byte(resp))
 								pp++
 							}
 						}

internal/translator/gemini/openai/responses/gemini_openai-responses_request.go

@@ -143,17 +143,11 @@ func ConvertOpenAIResponsesRequestToGemini(modelName string, inputRawJSON []byte
 				}
 
 				functionResponse, _ = sjson.Set(functionResponse, "functionResponse.name", functionName)
-				// Also set response.name to align with docs/convert-2.md
-				functionResponse, _ = sjson.Set(functionResponse, "functionResponse.response.name", functionName)
 
 				// Parse output JSON string and set as response content
 				if output != "" {
 					outputResult := gjson.Parse(output)
-					if outputResult.IsObject() {
-						functionResponse, _ = sjson.SetRaw(functionResponse, "functionResponse.response.content", outputResult.String())
-					} else {
-						functionResponse, _ = sjson.Set(functionResponse, "functionResponse.response.content", output)
-					}
+					functionResponse, _ = sjson.Set(functionResponse, "functionResponse.response.result", outputResult.Raw)
 				}
 
 				functionContent, _ = sjson.SetRaw(functionContent, "parts.-1", functionResponse)

internal/translator/openai/claude/openai_claude_request.go

@@ -133,27 +133,16 @@ func ConvertClaudeRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 					return true
 				})
 
-				// Create main message if there's text content or tool calls
-				if len(contentItems) > 0 || len(toolCalls) > 0 {
+				// Emit text/image content as one message
+				if len(contentItems) > 0 {
 					msgJSON := `{"role":"","content":""}`
 					msgJSON, _ = sjson.Set(msgJSON, "role", role)
 
-					// Set content
-					if len(contentItems) > 0 {
-						contentArrayJSON := "[]"
-						for _, contentItem := range contentItems {
-							contentArrayJSON, _ = sjson.SetRaw(contentArrayJSON, "-1", contentItem)
-						}
-						msgJSON, _ = sjson.SetRaw(msgJSON, "content", contentArrayJSON)
-					} else {
-						msgJSON, _ = sjson.Set(msgJSON, "content", "")
-					}
-
-					// Set tool calls for assistant messages
-					if role == "assistant" && len(toolCalls) > 0 {
-						toolCallsJSON, _ := json.Marshal(toolCalls)
-						msgJSON, _ = sjson.SetRaw(msgJSON, "tool_calls", string(toolCallsJSON))
+					contentArrayJSON := "[]"
+					for _, contentItem := range contentItems {
+						contentArrayJSON, _ = sjson.SetRaw(contentArrayJSON, "-1", contentItem)
 					}
+					msgJSON, _ = sjson.SetRaw(msgJSON, "content", contentArrayJSON)
 
 					contentValue := gjson.Get(msgJSON, "content")
 					hasContent := false
@@ -168,11 +157,19 @@ func ConvertClaudeRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 						hasContent = contentValue.Raw != "" && contentValue.Raw != "null"
 					}
 
-					if hasContent || len(toolCalls) != 0 {
+					if hasContent {
 						messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(msgJSON).Value())
 					}
 				}
 
+				// Emit tool calls in a separate assistant message
+				if role == "assistant" && len(toolCalls) > 0 {
+					toolCallMsgJSON := `{"role":"assistant","tool_calls":[]}`
+					toolCallsJSON, _ := json.Marshal(toolCalls)
+					toolCallMsgJSON, _ = sjson.SetRaw(toolCallMsgJSON, "tool_calls", string(toolCallsJSON))
+					messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(toolCallMsgJSON).Value())
+				}
+
 			} else if contentResult.Exists() && contentResult.Type == gjson.String {
 				// Simple string content
 				msgJSON := `{"role":"","content":""}`

internal/translator/openai/openai/responses/openai_openai-responses_request.go

@@ -2,6 +2,7 @@ package responses
 
 import (
 	"bytes"
+
 	"github.com/tidwall/gjson"
 	"github.com/tidwall/sjson"
 )
@@ -147,6 +148,11 @@ func ConvertOpenAIResponsesRequestToOpenAIChatCompletions(modelName string, inpu
 
 			return true
 		})
+	} else if input.Type == gjson.String {
+		msg := "{}"
+		msg, _ = sjson.Set(msg, "role", "user")
+		msg, _ = sjson.Set(msg, "content", input.String())
+		out, _ = sjson.SetRaw(out, "messages.-1", msg)
 	}
 
 	// Convert tools from responses format to chat completions format