OpenAI API Error: 405 Method not allowed. Only POST requests are supported. ???

[SkybuckFlying]

What happened?

# Set OpenAI-compatible variables
$env:OPENAI_API_KEY = "key123"
$env:OPENAI_BASE_URL = "http://api.together.xyz/v1"
$env:OPENAI_MODEL = "Qwen/Qwen3-235B-A22B-Instruct-2507-tput"

$env:GEMINI_API_KEY = "key123"  # Must be a non-empty string
$env:GEMINI_BASE_URL = "http://api.together.xyz/v1"
$env:GEMINI_MODEL = "Qwen/Qwen3-235B-A22B-Instruct-2507-tput"

# Set Gemini-compatible variables (for internal redirecting)
# $env:GEMINI_API_KEY = "dummy"  # Must be a non-empty string
# $env:GEMINI_BASE_URL = "http://127.0.0.1:1234/v1"
# $env:GEMINI_MODEL = "google/gemma-3n-e4b"

# Optional: run your main script
qwen.ps1

I tried with and without gemini settings.

What is wrong ? how to do it right ?:

(I do have key):

│ ⚠  Failed to load tiktoken, falling back to character approximation: Error: Missing tiktoken_bg.wasm            │
│        at node_modules/tiktoken/tiktoken.cjs                                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154029:30)                   │
│        at __require2 (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:19:51)     │
│        at file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154382:79              │
│        at async OpenAIContentGenerator.countTokens                                                              │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154382:36)                   │
│        at async GeminiClient.tryCompressChat                                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247415:49)                   │
│        at async GeminiClient.sendMessageStream                                                                  │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247222:24)                   │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255622:24        │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255715:34        │
│ ⚠  Failed to load tiktoken, falling back to character approximation: TypeError: get_encoding is not a function  │
│        at OpenAIContentGenerator.countTokens                                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154383:28)                   │
│        at async GeminiClient.sendMessageStream                                                                  │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247239:51)                   │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255622:24        │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255715:34        │
│ ⚠  Failed to load tiktoken, falling back to character approximation: TypeError: get_encoding is not a function  │
│        at OpenAIContentGenerator.countTokens                                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154383:28)                   │
│        at process.processTicksAndRejections (node:internal/process/task_queues:105:5)                           │
│        at async OpenAIContentGenerator.generateContentStream                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154283:38)                   │
│        at async retryWithBackoff                                                                                │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:246355:14)                   │
│        at async GeminiChat.sendMessageStream                                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:246735:31)                   │
│        at async Turn.run                                                                                        │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:245696:30)                   │
│        at async GeminiClient.sendMessageStream                                                                  │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247257:22)                   │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255622:24        │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255715:34        │
│ ✖  OpenAI API Streaming Error: 405 Method not allowed. Only POST requests are supported.                        │
│ ✖  Error when talking to Gemini API Full report available at:                                                   │
│    G:\Temp\gemini-client-error-Turn.run-sendMessageStream-2025-07-31T23-14-41-147Z.json                         │
│ ⚠  Failed to load tiktoken, falling back to character approximation: TypeError: get_encoding is not a function  │
│        at OpenAIContentGenerator.countTokens                                                                    │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154383:28)                   │
│        at process.processTicksAndRejections (node:internal/process/task_queues:105:5)                           │
│        at async OpenAIContentGenerator.generateContent                                                          │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:154136:38)                   │
│        at async retryWithBackoff                                                                                │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:246355:14)                   │
│        at async GeminiClient.generateJson                                                                       │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247300:22)                   │
│        at async checkNextSpeaker                                                                                │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:246254:28)                   │
│        at async GeminiClient.sendMessageStream                                                                  │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247269:32)                   │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255622:24        │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255715:34        │
│ ✖  OpenAI API Error: 405 Method not allowed. Only POST requests are supported.                                  │
│ ✖  Error generating JSON content via API. Full report available at:                                             │
│    G:\Temp\gemini-client-error-generateJson-api-2025-07-31T23-14-41-236Z.json                                   │
│ ⚠  Failed to talk to Gemini endpoint when seeing if conversation should continue. Error: Failed to generate     │
│    JSON content: OpenAI API error: 405 Method not allowed. Only POST requests are supported.                    │
│        at GeminiClient.generateJson                                                                             │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247345:13)                   │
│        at async checkNextSpeaker                                                                                │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:246254:28)                   │
│        at async GeminiClient.sendMessageStream                                                                  │
│    (file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:247269:32)                   │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255622:24        │
│        at async file:///G:/Tools/npm-global/node_modules/@qwen-code/qwen-code/bundle/gemini.js:255715:34   

What did you expect to happen?

It should work ?

Client information

│ CLI Version 0.0.1-alpha.12 │
│ Git Commit d360b86 (local modifications) │
│ Model Qwen/Qwen3-235B-A22B-Instruct-2507-tput │
│ Sandbox no sandbox │
│ OS win32 │
│ Auth Method openai

Login information

No response

Anything else we need to know?

No response

[hebra]

Same issue here. This issue seems to have started with the latest alpha.12 release.
Furthermore, this only seems to be an issue in environment with restricted network access.
E.g. in my case Qwen can only access AWS Bedrock via a LiteLLM proxy without any external network access to the wider Internet. In a different environment with the same set up but will Internet access the issue does not occur and qwen-code can access the Bedrock models and perform queries.

[hebra]

Fix for this on this PR #169

@SkybuckFlying are you able to check if my fix is resolving your issue?

[pomelo-nwu]

This issue has been open for over 6 months with no recent activity. It appears to be related to an early version of Qwen Code.
If you can still reproduce this issue with the latest version, please provide updated reproduction steps and we'll investigate.
Closing for now - feel free to reopen with fresh information.