[FEATURE]: OpenRouter Service Tiers Support for Reducing Model Cost

[landon-thull]

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

Overview

Implement configuration options for the OpenRouter service_tier parameter (see Related Documentation). This parameter provides access to reduced cost (0.5x) "flex" endpoints at the cost of increased latency. Currently, only OpenAI and Google models support this parameter.

Benefits

Reduced token cost for low-priority tasks on supported models.

Implementation Proposals

1. Hotkey Configuration in TUI (more complex)

Implement a service tier component into the status bar of the OpenCode TUI, similar to the thinking level component. Add support for rapid toggling via a hotkey.

Higher complexity, more user-friendly

2. Configuration File Option (less complex)

Add a default service tier configuration option that will be applied to models that support it.

Related Documentation

• OpenRouter Service Tiers
• Gemini Flex Inference

[github-actions]

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

• [FEATURE]: Cost calculation according to service_tier (OpenAI) #11597: Feature request for cost calculation based on service_tier (OpenAI) — closed
• [FEATURE]: Support OpenAI service tier (flex/priority processing) #7511: Feature request to support OpenAI service tier (flex/priority processing) — closed
• [FEATURE]: Support configuring OpenAI service_tier='flex' for chat/completions and document in config #12297: Feature request to support configuring OpenAI service_tier='flex' — closed
• [FEATURE]: Amazon Bedrock service tier support #25409: Similar feature request for service tier support (Amazon Bedrock) — open

Please review the above to see if your request is already covered or if this extends the prior work.

---

For the hotkey/TUI toggle aspect of your proposal, please also check our pinned keybinds documentation: #4997

[landon-thull]

Reviewed existing issues. They are similar, but they do not include any consideration for Google model support through OpenRouter.

I attempted to utilize the configuration option as shown in #12297, but OpenRouter logs still show "standard" service tier:

"provider": {
    "openrouter": {
      "models": {
        "google/gemini-3.5-flash": {
          "options": {
            "serviceTier": "flex"
          }
        }
      }
    }
  }

OpenRouter logs:

 "service_tier": "standard",
  "api_type": "completions",

[Moicky]

Yes this would be awesome! 👍

[Moicky]

@landon-thull I believe this can already be accomplished. I looked through some similar gh issues, and found a way.

OpenRouter Generation Info:

My config:

 "provider": {
    "openrouter": {
      "models": {
        "openai/gpt-5.5": {
          "options": {
            "serviceTier": "flex",
            "service_tier": "flex"
          }
        },
        "gpt-5.5": {
          "options": {
            "serviceTier": "flex",
            "service_tier": "flex"
          }
        }
      }
    }
  }

Not sure which of these worked, but at the end I get flex 😆