[FEATURE]: add GPT-5.4 fast mode (/fast) as a model control

[Pribess]

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

Add GPT-5.4 fast mode (/fast) to the TUI.

This feature was introduced with GPT-5.4, and opencode should expose it in a first-class way.

User-facing behavior:

• /fast toggles fast mode
• the prompt status line shows fast when enabled
• supported models send the fast-mode request option

Implementation direction:

• represent fast support as model-level controls metadata
• keep support checks, UI state, and request behavior in sync
• keep legacy fast fields as compatibility aliases during migration

[github-actions]

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

• [FEATURE]: Support for Claude Opus 4.6 Fast Mode #12923: Similar feature request for fast mode support (Claude Opus 4.6 Fast Mode) — the underlying feature ask (exposing fast mode as a first-class model control in OpenCode) is essentially the same, just for a different model.

[zedd3v]

@rekram1-node @thdxr

[zedd3v]

bump

[flmalte]

Is there any way to manually enable it?

[IIIIOvOIIII]

I second this issue.

[yednc]

我也有同样的问题

[bxb1337]

it will be great

[Borber]

I also really want to try this feature.

[bxb1337]

Is there any way to manually enable it?

manually override in your opencode.json

        "gpt-5.3-codex": {
          "name": "GPT-5.3 Codex",
          "limit": {
            "context": 400000,
            "output": 128000
          },
          "options": {
            "serviceTier": "priority"
          },
          "modalities": {
            "input": [
              "text",
              "image",
              "pdf"
            ],
            "output": [
              "text"
            ]
          },
          "variants": {
            "xhigh": {
              "reasoningEffort": "xhigh",
              "textVerbosity": "low",
            },
            "high": {
              "reasoningEffort": "high",
              "textVerbosity": "low",
            },
            "medium": {
              "reasoningEffort": "medium",
              "textVerbosity": "low",
            },
            "low": {
              "reasoningEffort": "low",
              "textVerbosity": "low",
            },
            "none": {
              "reasoningEffort": "none",
              "textVerbosity": "low",
            }
          },
        }

[josevelaz]

Is there any way to manually enable it?

manually override in your opencode.json

        "gpt-5.3-codex": {
          "name": "GPT-5.3 Codex",
          "limit": {
            "context": 400000,
            "output": 128000
          },
          "options": {
            "serviceTier": "priority"
          },
          "modalities": {
            "input": [
              "text",
              "image",
              "pdf"
            ],
            "output": [
              "text"
            ]
          },
          "variants": {
            "xhigh": {
              "reasoningEffort": "xhigh",
              "textVerbosity": "low",
            },
            "high": {
              "reasoningEffort": "high",
              "textVerbosity": "low",
            },
            "medium": {
              "reasoningEffort": "medium",
              "textVerbosity": "low",
            },
            "low": {
              "reasoningEffort": "low",
              "textVerbosity": "low",
            },
            "none": {
              "reasoningEffort": "none",
              "textVerbosity": "low",
            }
          },
        }

Could you share a comparison on how fast mode is supposed to look?
I added the property to the options but I'm not seeing any difference I don't think