Model does not support image input Codex

Quick fix

Add an image-input modality declaration for GLM-5.3-Flash to avoid inheriting a model configuration containing only ["text"].

Symptom

Error outputCopy

Codex 直接拦截图片输入：「此模型不支持图片输入」。

English translation

Codex directly blocks image input: “This model does not support image input.”

Cause

The Zhipu GLM preset modelCatalog does not include GLM-5.3-Flash. Existing seed models explicitly declare `inputModalities: ["text"]`. After changing the model to GLM-5.3-Flash, the hidden field inherits this text-only declaration and saves it to the provider's modelCatalog.

When generating the Codex catalog, an explicit text-only declaration takes precedence over fail-open behavior. Therefore, `glm-5.3-flash` is marked as not supporting image input. Without this incorrect explicit declaration, unknown models fail open to `["text", "image"]`.

Fix

1. 1

   Add `glm-5.3-flash` to the Codex preset for Zhipu GLM and explicitly set its input modalities to `["text", "image"]`; the issue has a fix branch prepared with this change and regression tests.

   src/config/codexProviderPresets.ts

   model: "glm-5.3-flash"
   inputModalities: ["text", "image"]

Affected Versions

ToolCodex

Version受影响：3.20.4

PlatformsmacOS

Source Issues

This page is distilled from 1 real issues

• #7688Zhipu GLM 预设缺少 glm-5.3-flash 且文本模态声明会覆盖 fail-open，导致 Codex 提示「此模型不支持图片输入」

FAQ

(No FAQ yet)

Related problems

Only under the OpenCode agentUpdate to v3.16.4 or later to use OpenCode Go presets with API Key-only configuration in Claude Code, Claude Desktop, Codex, and OpenCode.1 issues merged401 UnauthorizedUpgrade to cc-switch 4.0.x, or clear the stale third-party OPENAI_API_KEY from ~/.codex/auth.json manually, restart Codex, and sign in again with ChatGPT OAuth.2 issues mergedThe model cannot discover any tools - CodexUpgrade cc-switch past 3.16.1 and edit cc-switch-model-catalog.json to set supports_search_tool to false for the DeepSeek models.3 issues merged