OpenCode Go — API Proxy Returns cost: "0" and Dashboard Not Tracking Usage #26213

Description

Summary

External API calls to OpenCode Go (https://opencode.ai/zen/go/v1/chat/completions) return "cost": "0" in the response payload for all models tested. The usage dashboard at opencode.ai shows inconsistent or missing tracking for these calls, while the requests themselves complete successfully.

Environment

• Provider: OpenCode Go (flat $10/mo subscription)
• Endpoint: https://opencode.ai/zen/go/v1/chat/completions
• Auth: Bearer API key (standard Go subscription key)
• Models tested: deepseek-v4-flash, mimo-v2.5-pro, qwen3.6-plus
• Client: (OpenAI-compatible HTTP client), also reproduced with bare curl
• OS: WSL2 (Ubuntu) on Windows 11

Steps to Reproduce

1. Send a POST request to https://opencode.ai/zen/go/v1/chat/completions with a valid Go API key and any Go model ID.
2. Observe the "cost" field in the response.
3. Check the Go usage dashboard at opencode.ai.

Example request:
curl -s -X POST "https://opencode.ai/zen/go/v1/chat/completions"
-H "Authorization: Bearer $OPENCODE_GO_API_KEY"
-H "Content-Type: application/json"
-d '{"model":"deepseek-v4-flash","messages":[{"role":"user","content":"say OK"}],"max_tokens":50}'

Expected Behavior

• Response "cost" field reflects the actual token cost for the request.
• Dashboard updates to reflect usage against the rolling ($12/5hr), weekly ($30), and monthly ($60) limits.

Actual Behavior

• Response always returns "cost": "0" across all models tested, regardless of token count.
• Dashboard tracking is inconsistent:
  • Cron-job-style direct API calls appear with ~15–20 minute delay.
  • Interactive session calls via non-TUI clients show significantly longer delays or do not appear at all.
  • No OpenRouter or Gemini usage is leaking into Go — calls are correctly routed.

Plugins

No response

OpenCode version

No response

Steps to reproduce

No response

Screenshot and/or share link

No response

Operating System

No response

Terminal

No response