Failed to fetch models from Ollama even though ollama is running; Github Copilot Chat does not filter out Ollama Cloud Models

[AadityaRavindran]

Does this issue occur when all extensions are disabled?: Yes (even if GitHub Copilot Chat is the only extension enabled)

I have ollama setup to use models locally, but vscode (snap), vscode-insiders (snap and apt variants) are unable to detect any models when ollama cloud models are installed. The continue.dev extension however works fine and is able to list ollama models with the autodetect defaults. I even tried disabling all extensions except for copilot chat. But I see the same issue. Also, the copilot extension is able to detect all cloud models I have access to on my github account.

• GitHub Copilot Chat Extension: 0.34.2025112701 (prod)
• VS Code: 1.107.0-insider (c9b7a5a)
• OS: linux 6.14.0-1016-oem x64
• Ollama version: 0.12.10

User Settings:

  "http.systemCertificatesNode": false,
  "github.copilot.advanced.debug.useElectronFetcher": true,
  "github.copilot.advanced.debug.useNodeFetcher": false,
  "github.copilot.advanced.debug.useNodeFetchFetcher": true

2025-11-27 12:15:46.801 [error] Error: Failed to fetch models from Ollama. Please ensure Ollama is running. If ollama is on another host, please configure the `"github.copilot.chat.byok.ollamaEndpoint"` setting.
	at IS.getAllModels (/home/user/.vscode-insiders/extensions/github.copilot-chat-0.34.2025112701/dist/extension.js:1220:11553)
	at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
	at async IS.provideLanguageModelChatInformation (/home/user/.vscode-insiders/extensions/github.copilot-chat-0.34.2025112701/dist/extension.js:1220:8846)
	at async wA.$provideLanguageModelChatInfo (file:///usr/share/code-insiders/resources/app/out/vs/workbench/api/node/extensionHostProcess.js:118:30901): Error fetching available Ollama models
2025-11-27 12:15:46.812 [error] Error: Failed to fetch models from Ollama. Please ensure Ollama is running. If ollama is on another host, please configure the `"github.copilot.chat.byok.ollamaEndpoint"` setting.
	at IS.getAllModels (/home/user/.vscode-insiders/extensions/github.copilot-chat-0.34.2025112701/dist/extension.js:1220:11553)
	at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
	at async IS.provideLanguageModelChatInformation (/home/user/.vscode-insiders/extensions/github.copilot-chat-0.34.2025112701/dist/extension.js:1220:8846)
	at async wA.$provideLanguageModelChatInfo (file:///usr/share/code-insiders/resources/app/out/vs/workbench/api/node/extensionHostProcess.js:118:30901): Error fetching available Ollama models
2025-11-27 12:15:56.883 [debug] [context keys] Window state change. Needs offline check: false, active: true, focused: false.
2025-11-27 12:16:19.001 [debug] [context keys] Window state change. Needs offline check: false, active: true, focused: true.

Ollama endpoint settings are defaulted - "github.copilot.chat.byok.ollamaEndpoint": http://localhost:11434

user@host:~/$ curl http://localhost:11434
Ollama is running
user@host:~/$ ollama ls
gpt-oss:20b                            17052f91a42e    13 GB     3 days ago     
MFDoom/deepseek-r1-tool-calling:70b    e0c424928d2a    42 GB     3 days ago     
glm4:9b                                5b699761eca5    5.5 GB    3 days ago     
gemini-3-pro-preview:latest            91a1db042ba1    -         3 days ago     
nate/instinct:latest                   aee59dcab680    4.7 GB    2 weeks ago    
nomic-embed-text:latest                0a109f422b47    274 MB    2 weeks ago    
qwen2.5-coder:7b                       dae161e27b0e    4.7 GB    2 weeks ago    
gpt-oss:120b                           a951a23b46a1    65 GB     2 weeks ago    
deepseek-r1:70b                        d37b54d01a76    42 GB     2 weeks ago    
qwen3-coder:30b                        06c1097efce0    18 GB     2 weeks ago
...

Using Network Proxy Test shows that it should have worked fine

Proxy:
- URL: None (0 ms)

Sending GET request to http://localhost:11434...
Received response:
- Status: 200 OK

Sending GET request to http://localhost:11434 using fetch from Electron...
Received response:
- Status: 200 OK

Sending GET request to http://localhost:11434 using fetch from Node.js...
Received response:
- Status: 200 OK

Sending GET request to http://localhost:11434 using fetch from Node.js (allow HTTP2)...
Received response:
- Status: 200 OK

Possible duplicates: #279407 (but that one shows a different error log and is on a different os)

Steps to Reproduce:

1. Pull any Ollama cloud model (Eg: gemini-3-pro-preview:latest)
2. Open VSCode/Insiders -> Go to Chat -> Manage Models
3. Add Models -> Select Ollama. See that no ollama models show up.

Update: When cloud models like gemini-3-pro-preview:latest were removed from ollama, Github Copilot Chat started listing all the models.

[lramos15]

This isn't a remote scenario, right? It's just Ollama running on a local linux box with VS code running on the same linux box?

[AadityaRavindran]

This isn't a remote scenario, right? It's just Ollama running on a local linux box with VS code running on the same linux box?

Correct. I'm running everything on one machine. No wsl or virtual box either. I'm running ubuntu 24.04 natively.

[lramos15]

Very bizarre. Do you have an ollama base url configured? Does the full tags URL resolve in your VS code terminal? curl http://localhost:11434/api/tags

[AadityaRavindran]

curl http://localhost:11434/api/tags

Yes, I have a whole lot of models installed already. Like I said, continue.dev works fine.

{"models":[{"name":"gpt-oss:latest","model":"gpt-oss:latest","modified_at":"2025-11-27T12:32:23.156059143-05:00","size":13793441244,"digest":"17052f91a42e97930aa6e28a6c6c06a983e6a58dbb00434885a0cf5313e376f7","details":{"parent_model":"","format":"gguf","family":"gptoss","families":["gptoss"],"parameter_size":"20.9B","quantization_level":"MXFP4"}},{"name":"gpt-oss:20b","model":"gpt-oss:20b","modified_at":"2025-11-23T15:48:59.727847455-05:00","size":13793441244,"digest":"17052f91a42e97930aa6e28a6c6c06a983e6a58dbb00434885a0cf5313e376f7","details":{"parent_model":"","format":"gguf","family":"gptoss","families":["gptoss"],"parameter_size":"20.9B","quantization_level":"MXFP4"}},{"name":"MFDoom/deepseek-r1-tool-calling:70b","model":"MFDoom/deepseek-r1-tool-calling:70b","modified_at":"2025-11-23T15:43:15.727372989-05:00","size":42520399039,"digest":"e0c424928d2a19769439231d5ef23671c975130b003d6a52c66bb04b81b678c4","details":{"parent_model":"","format":"gguf","family":"llama","families":["llama"],"parameter_size":"70.6B","quantization_level":"Q4_K_M"}},{"name":"glm4:9b","model":"glm4:9b","modified_at":"2025-11-23T13:56:12.123585809-05:00","size":5455326235,"digest":"5b699761eca535dc55047ad9d2dbf54e3b8697709419ef78a70503ed4bfbcf44","details":{"parent_model":"","format":"gguf","family":"chatglm","families":["chatglm"],"parameter_size":"9.4B","quantization_level":"Q4_0"}},{"name":"gemini-3-pro-preview:latest","model":"gemini-3-pro-preview:latest","remote_model":"gemini-3-pro-preview","remote_host":"https://ollama.com:443","modified_at":"2025-11-23T13:41:39.545138338-05:00","size":363,"digest":"91a1db042ba1f0f926377ed682ec44c1dfff0395d646d45da21009f6cd319a43","details":{"parent_model":"","format":"","family":"","families":null,"parameter_size":"","quantization_level":""}},{"name":"nate/instinct:latest","model":"nate/instinct:latest","modified_at":"2025-11-09T18:02:15.762445102-05:00","size":4683090270,"digest":"aee59dcab68029d730d45b9d567abf034873074c22d6b6e25d63bfa1ec6a04e9","details":{"parent_model":"","format":"gguf","family":"qwen2","families":["qwen2"],"parameter_size":"7.6B","quantization_level":"Q4_K_M"}},{"name":"nomic-embed-text:latest","model":"nomic-embed-text:latest","modified_at":"2025-11-09T17:39:00.775812819-05:00","size":274302450,"digest":"0a109f422b47e3a30ba2b10eca18548e944e8a23073ee3f3e947efcf3c45e59f","details":{"parent_model":"","format":"gguf","family":"nomic-bert","families":["nomic-bert"],"parameter_size":"137M","quantization_level":"F16"}},{"name":"qwen2.5-coder:7b","model":"qwen2.5-coder:7b","modified_at":"2025-11-09T17:37:59.6561479-05:00","size":4683087561,"digest":"dae161e27b0e90dd1856c8bb3209201fd6736d8eb66298e75ed87571486f4364","details":{"parent_model":"","format":"gguf","family":"qwen2","families":["qwen2"],"parameter_size":"7.6B","quantization_level":"Q4_K_M"}},{"name":"gpt-oss:120b","model":"gpt-oss:120b","modified_at":"2025-11-09T17:14:55.569230074-05:00","size":65369818941,"digest":"a951a23b46a1f6093dafee2ea481d634b4e31ac720a8a16f3f91e04f5a40ecd9","details":{"parent_model":"","format":"gguf","family":"gptoss","families":["gptoss"],"parameter_size":"116.8B","quantization_level":"MXFP4"}},{"name":"deepseek-r1:70b","model":"deepseek-r1:70b","modified_at":"2025-11-09T17:01:10.822552427-05:00","size":42520397873,"digest":"d37b54d01a76512a5adbaa9da4ee1c00833bf2dfefb0d4cc54351ba529e0f4da","details":{"parent_model":"","format":"gguf","family":"llama","families":["llama"],"parameter_size":"70.6B","quantization_level":"Q4_K_M"}},{"name":"qwen3-coder:30b","model":"qwen3-coder:30b","modified_at":"2025-11-09T16:42:15.747287161-05:00","size":18556700761,"digest":"06c1097efce0431c2045fe7b2e5108366e43bee1b4603a7aded8f21689e90bca","details":{"parent_model":"","format":"gguf","family":"qwen3moe","families":["qwen3moe"],"parameter_size":"30.5B","quantization_level":"Q4_K_M"}},{"name":"gemma3:27b","model":"gemma3:27b","modified_at":"2025-11-09T16:41:38.346797464-05:00","size":17396936941,"digest":"a418f5838eaf7fe2cfe0a3046c8384b68ba43a4435542c942f9db00a5f342203","details":{"parent_model":"","format":"gguf","family":"gemma3","families":["gemma3"],"parameter_size":"27.4B","quantization_level":"Q4_K_M"}},{"name":"qwen3-vl:32b","model":"qwen3-vl:32b","modified_at":"2025-11-09T16:40:44.669091853-05:00","size":20910297800,"digest":"ff2e46876908853b8d0872595047c46d599d970056277e4864c8a1f0fd93fd3b","details":{"parent_model":"","format":"gguf","family":"qwen3vl","families":["qwen3vl"],"parameter_size":"33.4B","quantization_level":"Q4_K_M"}},{"name":"mistral:7b","model":"mistral:7b","modified_at":"2025-11-09T15:27:22.255609074-05:00","size":4372824384,"digest":"6577803aa9a036369e481d648a2baebb381ebc6e897f2bb9a766a2aa7bfbc1cf","details":{"parent_model":"","format":"gguf","family":"llama","families":["llama"],"parameter_size":"7.2B","quantization_level":"Q4_K_M"}},{"name":"embeddinggemma:300m","model":"embeddinggemma:300m","modified_at":"2025-11-09T15:16:26.924615721-05:00","size":621875917,"digest":"85462619ee721b466c5927d109d4cb765861907d5417b9109caebc4e614679f1","details":{"parent_model":"","format":"gguf","family":"gemma3","families":["gemma3"],"parameter_size":"307.58M","quantization_level":"BF16"}},{"name":"nomic-embed-text:137m-v1.5-fp16","model":"nomic-embed-text:137m-v1.5-fp16","modified_at":"2025-11-09T15:15:55.118675118-05:00","size":274302450,"digest":"0a109f422b47e3a30ba2b10eca18548e944e8a23073ee3f3e947efcf3c45e59f","details":{"parent_model":"","format":"gguf","family":"nomic-bert","families":["nomic-bert"],"parameter_size":"137M","quantization_level":"F16"}}]}

[lramos15]

Do you have an ollama base url configured?

This part. Do you have any settings? As the network proxy shows all URLs are working properly. And therefore the only thing I can think of is that we're hitting a bad URL. I can add some logs to the next insiders to also help in the future

[AadityaRavindran]

Do you have an ollama base url configured?

This part. Do you have any settings? As the network proxy shows all URLs are working properly. And therefore the only thing I can think of is that we're hitting a bad URL. I can add some logs to the next insiders to also help in the future

Yes. It should have been the default, but I also explicitly verified that "github.copilot.chat.byok.ollamaEndpoint": http://localhost:11434 for all my profiles.

[AadityaRavindran]

Logan Ramos (@lramos15) do you still need more info, of not, the label can go away right?

[TheXDS]

This seems to be pretty universal (at least if we're talking about Ollama not running in the same machine as VS Code)

Environment

• Windows 11 25H2
• VS Code 1.106.3 (user install), mostly all default settings throughout
• Ollama 0.13.1 hosted in remote server on the local network, accessible to other tools just fine (Continue.dev, Open WebUI, etc.)

Error message:

[error] Error: Failed to fetch models from Ollama. Please ensure Ollama is running. If ollama is on another host, please configure the `"github.copilot.chat.byok.ollamaEndpoint"` setting.
	at Vk.getAllModels (c:\Users\<my user>\.vscode\extensions\github.copilot-chat-0.33.5\dist\extension.js:1125:11553)
	at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
	at async Vk.provideLanguageModelChatInformation (c:\Users\<my user>\.vscode\extensions\github.copilot-chat-0.33.5\dist\extension.js:1125:8846)
	at async uA.$provideLanguageModelChatInfo (file:///c:/Users/<my user>/AppData/Local/Programs/Microsoft%20VS%20Code/resources/app/out/vs/workbench/api/node/extensionHostProcess.js:116:30605): Error fetching available Ollama models

Yes, I did configure the github.copilot.chat.byok.ollamaEndpoint to point to my Ollama server (in the format http://myollamaserver.mydomain.local:11434 (using the IP address also fails)... Could it be a formatting issue (cured by maybe checking for what the user enters and improving docs in that area)?

Yes, I did try it disabling pretty much anything that was not essential for VS Code and Copilot to work.

[WinPooh32]

Just updated to version 1.107.0-1765353552 and Copilot began to fail on fetching models from remote Ollama (0.13.1). The reason was models removed from disk. Ollama returned 404 when requesting such models in the list. Removing that models from the list fixed the issue.

If there are no connection logs at Ollama, issue can be in firewall, network configuration.

[AadityaRavindran]

Just updated to version 1.107.0-1765353552 and Copilot began to fail on fetching models from remote Ollama (0.13.1). The reason was models removed from disk. Ollama returned 404 when requesting such models in the list. Removing that models from the list fixed the issue.

If there are no connection logs at Ollama, issue can be in firewall, network configuration.

Oh, you're right. Ollama Cloud models weren't being identified correctly. Once I deleted my ollama cloud model gemini-3-pro-preview:latest, the tool was able to pick up on all the local models.