Skip to content

fix(provider): cap OUTPUT_TOKEN_MAX at 32K so usable() stays above zero #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
NormalDudeBro wants to merge 1 commit into
base: codemaxxxing
Choose a base branch
from
Open

fix(provider): cap OUTPUT_TOKEN_MAX at 32K so usable() stays above zero #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion packages/opencode/src/provider/transform.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@ function mimeToModality(mime: string): Modality | undefined {
return undefined
}

export const OUTPUT_TOKEN_MAX = Flag.OPENCODE_EXPERIMENTAL_OUTPUT_TOKEN_MAX || Infinity
export const OUTPUT_TOKEN_MAX = Flag.OPENCODE_EXPERIMENTAL_OUTPUT_TOKEN_MAX || 32_000

// Maps npm package to the key the AI SDK expects for providerOptions
function sdkKey(npm: string): string | undefined {
Expand Down
18 changes: 18 additions & 0 deletions packages/opencode/test/session/compaction.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -540,6 +540,24 @@ describe("session.compaction.isOverflow", () => {
),
)

it.live(
"BUG: no overflow when output limit equals context (grok-4.5 shape)",
provideTmpdirInstance(() =>
Effect.gen(function* () {
const compact = yield* SessionCompaction.Service
// grok-4.5: context 500K, output 500K, no limit.input
const model = createModel({ context: 500_000, output: 500_000 })

// 31K total — a normal early-conversation size. usable must be
// context - reservedOutput (capped at OUTPUT_TOKEN_MAX = 32K), i.e. 468K.
// Regression: OUTPUT_TOKEN_MAX defaulted to Infinity, making
// usable = 500K - 500K = 0 and overflowing on every single message.
const tokens = { input: 30_000, output: 1_000, reasoning: 0, cache: { read: 0, write: 0 } }
expect(yield* compact.isOverflow({ tokens, model })).toBe(false)
}),
),
)

it.live(
"returns false when model context limit is 0",
provideTmpdirInstance(() =>
Expand Down