opencode never executes any creation actions #11064

Activity

1. sjasoft

   added

   bugSomething isn't workingSomething isn't working

   on Jan 29, 2026
2. github-actions

   assigned

   rekram1-node

   on Jan 29, 2026
3. github-actions commented on Jan 29, 2026

   

   github-actionsBot

   on Jan 29, 2026 – with GitHub Actions

   Contributor

   More actions

   This issue might be a duplicate of existing issues. Please check:

   • Ollama (qwen2.5-coder): tool calls (edit/write) show as executed but no files are created/modified #7030: Ollama (qwen2.5-coder): tool calls (edit/write) show as executed but no files are created/modified
   • Opencode in linux cannot access tools or create files #729: Opencode in linux cannot access tools or create files

   Feel free to ignore if none of these address your specific case.
4. sjasoft commented on Jan 29, 2026

   

   sjasoft

   on Jan 29, 2026

   Author

   More actions

   Actually it seems to fail on only some models. Specifically I have tried vllm, ollama, lmstudio for local models with similar or worse results. However I just tried gpt-4o mini and that worked fine. This makes me think there is something amiss or finicky in the tool call flow for the models that did not work.

   Specifically I have tried the models in this opencode.json configuration.

   {
   "$schema": "https://opencode.ai/config.json",

   "default_agent": "build",
   "logLevel": "DEBUG",

   "model": "vllm/Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4",

   "plugin": [
   "opencode-antigravity-auth@latest"
   ],

   "permission": {
   "*": "allow"
   },

   "provider": {
   "lmstudio": {
   "npm": "@ai-sdk/openai-compatible",
   "name": "LM Studio (local)",
   "options": {
   "baseURL": "http://localhost:1234/v1",
   "timeout": 60000
   },
   "models": {
   "llama_3.x_8b_smarteaz_0.1b-i1": {
   "name": "llama_3.x_8b_smarteaz_0.1b-i1",
   "tools": true,
   "limit": {
   "context": 16384,
   "output": 4096
   }
   }
   }
   },

   "vllm": {
     "npm": "@ai-sdk/openai-compatible",
     "name": "vLLM (local)",
     "options": {
       "baseURL": "http://localhost:8000/v1",
       "headers": {
         "Authorization": "Bearer my-secret-key"
       }
     },
     "models": {
       "Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4": {
         "name": "Qwen/Qwen2.5-7B-Instruct-GPTQ-Int4",
         "tools": true,
         "limit": {
           "context": 32768,
           "output": 4096
         }
       },
       "Qwen/Qwen2.5-Coder-7B-Instruct-GPTQ-Int4": {
         "name": "Qwen/Qwen2.5-Coder-7B-Instruct-GPTQ-Int4",
         "tools": true,
         "limit": {
           "context": 32768,
           "output": 4096
         }
       }
     }
   },
   
   "google": {
     "models": {
       "antigravity-gemini-3-flash": {
         "name": "Gemini 3 Flash (Antigravity)",
         "limit": {
           "context": 1048576,
           "output": 65536
         },
         "modalities": {
           "input": [
             "text",
             "image",
             "pdf"
           ],
           "output": [
             "text"
           ]
         }
       }
     }
   }
   

   },

   "agent": {
   "build": {
   "model": "vllm/Qwen/Qwen2.5-Coder-7B-Instruct-GPTQ-Int4"
   }
   },

   "mcp": {
   "local-mcp": {
   "type": "local",
   "command": [
   "bun",
   "x",
   "opencode-mcp-local"
   ],
   "enabled": true
   }
   }
   }

   👀React with 👀1Reacted by Jaydon Chong
5. github-actions commented on Mar 30, 2026

   

   github-actionsBot

   on Mar 30, 2026 – with GitHub Actions

   Contributor

   More actions

   To stay organized issues are automatically closed after 90 days of no activity. If the issue is still relevant please open a new one.
6. github-actions

   closed this as completedon Mar 30, 2026
7. Jaydonchong commented on May 7, 2026

   

   Jaydonchong

   on May 7, 2026

   More actions

   Currently facing this issue as well, when asking to create a single file such as 'hello world' it is able to create it. However when asked to create a project (simple todolist) it somehow chooses to keep giving instructions rather than actually executing it.

   Usage: Ollama with qwen2.5-coder:3b

   👍React with 👍2Reacted by kk and Emil Sierżęga
8. kannandreams commented on Jun 14, 2026

   

   kannandreams

   on Jun 14, 2026

   More actions

   @Jaydonchong - I also faced the same issue.