- Added long-context pricing tiers and billing policies for subscription Codex models in the catalog.
- Introduced the `extendedContext` configuration setting to control premium long-context windows.
- Implemented runtime policy refresh and model re-binding when context settings change.
- Added comprehensive unit tests for pricing tiers, context capping, and policy toggling behavior.
- Replaced the `ctok` implementation with the `utok` universal tokenizer supporting multiple model families and UTF text encodings.
- Added tokenizer support and embedding data for Qwen3, DeepSeek V3, Kimi K2, and GLM-5 model variants.
- Added fixture generation scripts, vocabulary packers, and golden test suites for validating tokenization parity.
- Updated dependency requirements and Bazel workspace definitions for new crates and tools.
- Fixed `muse-spark-1.2` and `muse-spark-1.2-contributor` failing on tool-call turns by mapping them to `openai-responses` in `OPENCODE_GO_API_ID_OVERRIDES`.
- Added automatic fallback routing to borrow `openai-responses` routes from sibling gateways or billing-variant base IDs for gateway-first models.
- Configured `dropCachedModelIdsOnStaticMismatch` to ensure cached models holding stale static metadata are invalidated on discovery.
GMI Cloud's /v1/models returns only bare {id} rows, so dynamic discovery
resolved every model except the single bundled DeepSeek-V4-Flash seed with
a null context window, zero pricing, and no reasoning/thinking config.
Give the gmi-cloud mapper the same cross-provider canonical fallback that
SiliconFlow uses for the identical open-weight models: recover context
window, output limit, reasoning, and thinking ladder from the bundled
reference index while never borrowing another provider's pricing.
Fixes#8890
Cursor advertises Grok 4.5/4.6 as per-effort sibling ids
(cursor-grok-4.6-low|-medium|-high|-xhigh plus -fast variants), but
VARIANT_COLLAPSE_TABLES had no cursor entry, so the model hub showed 14
unrouted siblings instead of one logical model with effort routing.
GetUsableModels ships no thinkingDetails and the bundled references read
reasoning:false, so the picker also treated them as non-reasoning.
- Add CURSOR_VARIANT_COLLAPSE_TABLE folding each service-tier lane
(standard + -fast) into one logical model with effort routing onto the
live wire ids, mirroring Devin's grok-4-5 collapse.
- Rename the generic devinTierFamily/DevinTierRoutes helper to
tierFamily/TierRoutes now that both Devin and Cursor tables use it.
- Mark versioned cursor-grok-<version> ids as reasoning during discovery
(grok-code-* coding models stay non-reasoning).
Fixes#8803
GitHub Copilot serves grok-4.6 / grok-4.6-1m only via /responses, but
isCopilotResponsesModelId matched grok-4.5 exactly, so both the static
generator and dynamic discovery classified grok-4.6 as openai-completions
and requests 400d with unsupported_api_for_model.
- match grok-4.6 in isCopilotResponsesModelId
- add grok-4.6 / grok-4.6-1m to COPILOT_CACHE_INVALIDATED_MODEL_IDS so
stale cached completion routes drop on refresh
- regenerate github-copilot/grok-4.6 to api openai-responses (compat
block dropped, xhigh effort added by the responses policy)
- cover discovery routing and cache migration in tests
Fixes#8807
The OpenCode Go gateway serves muse-spark-1.2 and
muse-spark-1.2-contributor only at /zen/go/v1/responses, but the
/zen/go/v1/models discovery omits the provider.npm hint, so the
resolver fell through to openai-completions. The completions parser
then closed the stream without a finish_reason on every tool-call turn.
Pin both ids to openai-responses in OPENCODE_GO_API_RESOLUTION, mirroring
the existing deepseek-v4-flash override, and add a resolver regression.
Fixes#8957
The shared SQLite model cache wrapped every read/write in a blanket catch that swallowed unrecoverable SQLITE_CORRUPT*/SQLITE_NOTADB failures as best-effort misses, and getSharedDb cached the broken handle. A physically corrupt models.db therefore permanently disabled cached catalogs across processes: a successful live discovery could never overwrite the corrupt cache, so a runtime extension with no bundled catalog was stuck with only its bootstrap model.
On those unrecoverable codes the cache now self-heals: close the handle, quarantine models.db(+-wal/-shm) to models.db.corrupt-<ts>, recreate a fresh database, and retry the operation once. SQLITE_BUSY, permission, and unrelated errors keep their existing best-effort paths. The SQLITE_BUSY/corruption classifiers moved to @oh-my-pi/pi-utils so the credential store and model cache share one implementation.
Fixes#8867
- Kept the shared cursor/interaction-query module as the single handler and deleted the duplicate local implementation in cursor.ts
- Added the named webFetchRequestQuery approval case (field 9 is named under the regenerated proto)
- Preserved the deliberate no-fake-VM-success semantics for setupVmEnvironmentArgs (review of #8047)
- Updated the field-9 regression test to assert the named decode of the raw same-field reply, which also pins the LEN-prefix wire framing
- Added `reasoning_effort` kwarg and top-level support for Qwen 3.8+ templates.
- Introduced `qwenTemplateReasoningEffort` compatibility option and identity helpers.
- Enabled default reasoning enforcement and updated cache provider invalidation.
- Added comprehensive unit and compatibility test suites for Qwen reasoning dials.
Review follow-up: a safe `-wm` row previously kept backend-parsed capability metadata (no 1M floor, no daybreak pricing) while its synthesized plain listing was enriched — the same model reported two different contexts. Both listings now derive fallback window, the 1M floor, and daybreak cost from the canonical plain slug; unknown `-wm` SKUs and non-worker models keep their verbatim slug-derived metadata.
The authoritative-discovery test now resolves the configured `openai-codex/gpt-5.6-luna` through the real model resolver and asserts the exact bound id (and that an explicit `-wm` config still resolves verbatim) instead of only checking `ids.toContain`.
Codex backend discovery advertises worker-mode SKUs under a `-wm` suffix (gpt-5.6-luna-wm). Authoritative discovery replaced the bundled catalog and kept those slugs verbatim, so a configured `openai-codex/gpt-5.6-luna` vanished from the resolved catalog and the resolver's fuzzy fallback selected `-wm` instead — a route some ChatGPT accounts reject.
Model discovery now recognizes the `-wm` suffix: when the bundled Codex catalog ships the plain SKU, the `-wm` row is also registered under its plain id (re-derived so the 1M-window floor and daybreak pricing keyed on that slug still apply). Unknown `-wm` SKUs keep their authoritative verbatim slug and non-worker models are untouched, so distinct models and other providers are unaffected.
CoreWeave Serverless Inference (W&B Inference) is a reseller with a
rotating model menu, but its catalog entry omitted
dynamicModelsAuthoritative. Runtime /v1/models discovery therefore
merged into the frozen bundled slice instead of replacing it, so stale
ids (e.g. moonshotai/Kimi-K3, Kimi-K2.5) stayed selectable and 404 at
request time. Matches sibling resellers (baseten, gmi-cloud, aiand,
bedrock-mantle). Bundled models remain the offline/failure fallback via
the authoritativeFreshProviders gating in ModelRegistry.
When a collapsed Gemini 3.6/3.7 Flash family routes user minimal onto the
same Cloud Code Assist wire id as low, emit thinkingLevel LOW. Those -low
SKUs reject MINIMAL with HTTP 400.
Alibaba Token Plan advertises both dated DeepSeek V4 snapshots, but only
deepseek-v4-flash-0731 had an entry in ALIBABA_TOKEN_PLAN_DISCOVERED_MODEL_LIMITS.
deepseek-v4-pro-0813 is not in ALIBABA_TOKEN_PLAN_STATIC_MODELS either (only the
undated deepseek-v4-pro is), so it fell through to `contextWindow: null` /
`maxTokens: null` — the #7486 symptom, still live for this one id.
Reasoning already worked: the `normalizedId.startsWith("deepseek-v4")` branch
gives it reasoning: true and the high/max effort ladder. Only the limits were
missing, so this is a one-entry fix at 1M context / 384K output, matching both
deepseek-v4-pro and deepseek-v4-flash-0731.
Extends the existing discovery test to advertise the id and assert its limits
and thinking config. Verified the test fails without the source change
(contextWindow/maxTokens come back null) and passes with it.
Refs #8847
Cursor grok-4.6-xhigh stalled after a short "I'll fetch the page"
preamble because interaction_query frames (including proto field 9)
were dropped and the server waited until the 300s idle watchdog fired.
- Updated GPT-5.6 context window floor to 1,000,000 tokens across discovery, policies, and tests.
- Updated model configurations and pricing parameters in catalog models JSON.
Point xai and xai-oauth at grok-4.6, already in the bundled catalog.
Tests pin the default id in models.json and load picker fixtures from
the catalog so the next bump does not rot hardcoded name or cost.
Add grok-4.6 to the SuperGrok Responses effort allowlist so /model
can select low/medium/high/xhigh. Stale omitReasoningEffort cache
rows no longer hide the dial. max is omitted because api.x.ai 400s.
can1357