@oh-my-pi/pi-catalog 15.13.0 → 15.13.2

package/CHANGELOG.md

@@ -2,20 +2,161 @@
 
 ## [Unreleased]
 
+## [15.13.2] - 2026-06-15
+
+### Added
+
+- Added the `ToolCallSyntax` union and `FALLBACK_TOOL_SYNTAX` constant to `@oh-my-pi/pi-catalog/identity` (re-exported from `@oh-my-pi/pi-ai/grammar`).
+- Added `preferredToolSyntax(modelId)` to `@oh-my-pi/pi-catalog/identity`, resolving a model's native tool-call syntax affinity from its family token (Claude→`anthropic`, GLM→`glm`, Kimi→`kimi`, Qwen→`qwen3`, DeepSeek→`deepseek`, OpenAI/gpt-oss→`harmony`, else the `xml` fallback).
+- Added `flux-1-schnell-fp8` to the Fireworks serverless model catalog
+- Added `gpt-oss-20b` to the Fireworks model catalog
+- Added `qwen3-embedding-8b` to the Fireworks model catalog
+- Added `qwen3-reranker-8b` to the Fireworks model catalog
+- Added `Gemma 4 E2B IT` and `Gemma 4 E4B IT` to the Google model catalog
+- Added `qwen/qwen3-asr-flash` to the Zenmux model catalog
+- Added sparse `supportsTools` model metadata so providers can mark models that require in-band tool-call formatting.
+
+### Changed
+
+- Kept non-tool-capable Fireworks serverless models in discovery results and marked them with `supportsTools: false` for fallback-aware handling
+- Extended `modelFamilyToken(modelId)` to classify Claude/OpenAI ids the structured parser misses (older dated forms such as `claude-3-5-sonnet-20241022` and `gpt-4o`), returning `anthropic`/`openai` instead of an empty token.
+
+## [15.13.1] - 2026-06-15
+
 ### Added
 
 - Added `modelFamilyToken(modelId)` to `@oh-my-pi/pi-catalog/identity`: a coarse vendor-lineage token (`anthropic`/`openai`/`gemini`/`kimi`/…) for "are two models the same family?" comparisons, backed by `parseKnownModel` canonical-id normalization. Opaque and comparison-only; kind/variant collapsed onto the vendor token ([#2406](https://github.com/can1357/oh-my-pi/issues/2406))
+
+### Changed
+
+- Changed catalog metadata to update a model’s per-token pricing to input 0.09 and output 0.18
+- Changed the same cataloged model’s maximum token limit from 384000 to 65536
+
+### Fixed
+
+- Fixed MiniMax-M3 catalog context for `minimax` and `minimax-cn` to report the documented 1M long-context tier instead of the upstream 512K pricing boundary ([#2576](https://github.com/can1357/oh-my-pi/issues/2576)).
+- Fixed OpenCode Go MiMo catalog metadata so title generation and other tool-enabled calls omit unsupported `tool_choice` instead of triggering provider 400s ([#2509](https://github.com/can1357/oh-my-pi/issues/2509)).
+- Fixed OpenCode Go `kimi-k2.7-code` catalog metadata so resolve-gate requests use automatic tool selection instead of Moonshot-rejected forced `tool_choice` ([#2546](https://github.com/can1357/oh-my-pi/issues/2546)).
+- Fixed Anthropic compat for the `github-copilot` host so `supportsEagerToolInputStreaming` defaults to `false` there, matching the Copilot proxy which rejects the per-tool `eager_input_streaming` field ([#2558](https://github.com/can1357/oh-my-pi/issues/2558)).
+- Scoped vLLM model cache validity to the discovery base URL so changed endpoints refetch immediately, and bounded built-in vLLM discovery requests with a timeout.
+
+## [15.12.6] - 2026-06-14
+
+### Added
+
 - Added GLM-5.2 to the bundled zai (GLM Coding Plan) catalog as the selectable 1M served model.
+
+### Changed
+
+- Pinned zai `glm-5.2` to 1M context during catalog generation so endpoint discovery and older fallbacks cannot regress it to 200k.
+- Replaced the hand-maintained `zhipu-coding-plan` GLM reasoning allowlist and vision regex with a `parseGlmModel` family classifier in `identity/classify.ts` (variant + vision + version), surfaced as `isReasoningGlmModelId` / `isGlmVisionModelId`. Discovery now derives reasoning/vision capability from the GLM family instead of a per-id list, so newly-bumped integers (`glm-5.3`, `glm-6`, …) are covered automatically while `-flash`/`-preview` and the vision `…v` shape stay correctly classified.
+
+## [15.12.4] - 2026-06-13
+
+### Added
+
 - Added bundled Fireworks models `deepseek-v4-flash`, `kimi-k2.7-code`, `minimax-m2.5`, `minimax-m3`, `nemotron-3-ultra-nvfp4`, `qwen3.6-plus`, and `qwen3.7-plus`
 - Changed
+
+### Changed
+
+- Model `contextWindow`/`maxTokens` are now `number | null`; discovery emits `null` when a provider reports no limit, replacing the `222222`/`8888` (`UNK_CONTEXT_WINDOW`/`UNK_MAX_TOKENS`) sentinels (now removed). Bundled `models.json` unknown limits are `null`.
+- Changed the `github-copilot` model context window to `524288` tokens
+- Changed Fireworks model discovery to source the control-plane `List Models` API (`GET /v1/accounts/fireworks/models?filter=supports_serverless=true`) instead of the OpenAI-compatible `/v1/models` inference listing. The inference endpoint returns a sparse, account-specific subset that omits on-demand serverless models (e.g. `kimi-k2.7-code`), so newly published serverless models stayed invisible in the picker until hand-added to the bundled catalog. The control-plane catalog enumerates every serverless model with capability metadata (`supportsServerless`/`supportsTools`/`supportsImageInput`/`contextLength`/`displayName`), paginated and filtered to tool-capable `READY` entries, then merged with bundled/models.dev references — the Kimi K2 max-output clamp and DeepSeek V4 thinking-toggle strip are preserved, and unbundled models default to reasoning so `buildModel` derives the Fireworks effort map. New serverless releases now surface automatically with no catalog edits.
+
+### Fixed
+
+- Filled missing `contextWindow` and `maxTokens` in generated `models.json` for proxy/reseller variants by inheriting limits from canonical-family and segment-reference models
+- Ignored zero-cost `x-ai` subscription entries as reference sources when backfilling limits so inflated values are not propagated
+- Fixed the model cache opening with `PRAGMA journal_mode=WAL` before `PRAGMA busy_timeout`, so concurrent omp startups could crash inside `getDb()` on `SQLITE_BUSY` during WAL recovery instead of waiting through the transient lock. The busy handler is now installed before the first lock-taking statement ([#2421](https://github.com/can1357/oh-my-pi/issues/2421)).
+
+## [15.11.8] - 2026-06-12
+
+### Fixed
+
+- Fixed Antigravity `gemini-3.1-pro --thinking high` failing with `Cloud Code Assist API error (400): Request contains an invalid argument.` — the upstream `gemini-3.1-pro-high` deployment rejects every `streamGenerateContent` request on both CCA endpoints while discovery still advertises it. High effort now routes to `gemini-pro-agent` (the same "Gemini 3.1 Pro (High)" model, verified accepting the identical request body), and the model-cache fingerprint version was bumped (`merge-v2` → `merge-v3`) so existing fresh caches refetch discovery and pick up the corrected routing immediately.
+
+## [15.11.7] - 2026-06-12
+
+### Added
+
 - Added effort-tier variant collapsing (`variant-collapse`): providers that expose one logical model as several effort/thinking-suffixed upstream ids (Antigravity CCA `gemini-3.5-flash-extra-low`/`-low`/`gemini-3-flash-agent`, `gemini-3[.1]-pro-low|high`, `claude-*[-thinking]` pairs, `gpt-oss-120b-medium`) collapse into one logical entry carrying per-effort upstream routing in `thinking.effortRouting` (plus `thinking.suppressWhenOff` for Cloud Code Assist ids whose baked server default re-applies when `thinkingConfig` is omitted). Request-time code resolves the outbound id via `resolveWireModelId(model, effort)`; selection, caching, and usage attribution key on the logical id.
 - Added the automatic `X`/`X-thinking` pair rule (`deriveThinkingPairFamilies`): any provider's live bare/thinking twin collapses into the bare id, routing thinking-enabled requests to the `-thinking` backing id (trailing or infix token, so `kimi-k2-thinking-turbo` pairs with `kimi-k2-turbo`). Gated on same api and compatible pricing — all-zero cost rows count as unknown, while twins that both carry real, differing prices remain separate SKUs.
 - Added `collapseBuiltModelVariants` and wired collapsing at every materialization point — Antigravity discovery, the catalog generator, and the model-manager merge — so stale sources (old static beside collapsed dynamic results, mixed cache rows) converge on logical entries instead of unioning raw tier ids back into the catalog.
 - Added `thinking.requiresEffort`, baked for reasoning-only upstreams — Gemini 3.x (levels only, no off), Gemini 2.5 Pro (thinkingBudget floors at 128, rejects 0), OpenAI o-series, MiniMax M2, and thinking-variant SKUs (`*-thinking`/`*-reasoner`/`*-reasoning`, with a negation-aware token grammar so `non-thinking` ids never match). Identity derivation bakes it for new entries and `fillThinkingWireDefaults` backfills explicit/cached metadata; `minimumSupportedEffort` exposes the canonical floor. Pair-collapsed twins drop member flags (their off routes to the bare SKU), while identity re-flags pairs whose logical id is itself mandatory
+
+### Changed
+
+- Changed model display names to drop model-extrinsic decorations: gateway author prefixes (`OpenAI: …`, `Google: …`), `(latest)` alias markers, `(Antigravity)` provider attribution, price tiers (`($$$$)`), and promo/lifecycle tags (`(20% off)`, `(retires …)`). `cleanModelName` is applied in `buildModel` (covers live discovery and stale caches) and as a catalog-generator pass; Antigravity discovery no longer appends `(Antigravity)` to display names. Variant tags that map to distinct wire ids (`(Thinking)`, `(free)`, `(Fast)`, dates, regions) are preserved.
+- Changed the `google-antigravity` default model from `gemini-3-pro-high` to `gemini-3.1-pro`
+- Changed `gemini-2.5-flash-thinking` handling from discovery-denylist to collapsing into `gemini-2.5-flash` (thinking-enabled requests route to the `-thinking` backing id)
+- Bumped the model cache schema to v5 so rows predating effort-tier variant collapsing (raw `-low`/`-high`/`-thinking` member ids) are invalidated
+
+### Fixed
+
+- Fixed catalog generation to apply effort-tier variant collapsing before provider grouping to ensure collapsed model families are consistently materialized without being impacted by in-loop mutation
+- Fixed Kimi K2.6 OpenAI-compatible compat metadata to use a 300s stream watchdog floor, covering Fire Pass router ids as well as public `kimi-k2.6` ids so long reasoning starts do not hit the generic first-event timeout ([#2366](https://github.com/can1357/oh-my-pi/issues/2366)).
+
+## [15.11.4] - 2026-06-12
+
+### Fixed
+
+- Fixed MiniMax M2-family and OpenAI gpt-oss model metadata so OpenAI-compatible catalog entries declare only `low|medium|high` thinking efforts. Their upstreams reject `minimal`, `xhigh`, and Fireworks' `minimal → none` wire mapping, so `fireworks/minimax-m2.7` as the smol auto-thinking classifier model 400ed on every turn. OpenAI-compatible provider effort maps (`Groq qwen/qwen3-32b`, DeepSeek-family, OpenRouter Anthropic adaptive, Fireworks `minimal → none`) now bake into `thinking.effortMap` in catalog metadata instead of `buildOpenAICompat`, and request builders read that field directly. Regenerated `models.json` now makes `disableReasoning` choose `low` for those families while leaving GLM-5.x and other Fireworks models on the existing `minimal → none` path ([#2315](https://github.com/can1357/oh-my-pi/issues/2315)).
+
+### Added
+
 - Added `requiresJuiceZeroHack` Responses-API compat flag, resolved by `buildOpenAIResponsesCompat` from GPT-5-family model names and overridable via sparse model `compat` config. Replaces the request-time `model.name.startsWith("gpt-5")` sniff that gated the trailing `# Juice: 0 !important` no-reasoning developer item.
+
+## [15.11.3] - 2026-06-11
+
+### Added
+
 - Added `requestModelId` on `Model` to represent the upstream model id used when a catalog entry is a local variant
 - Added synthetic GitHub Copilot long-context model variants with `-1m` suffixes when tiered token pricing is advertised
+
+### Changed
+
+- Changed GitHub Copilot discovery to request `X-GitHub-Api-Version: 2026-06-01` from `api.githubcopilot.com`
+- Changed GitHub Copilot discovery to cap base model `contextWindow` to the default token tier and keep long-context access as the separate `-1m` model entry
+- Changed Copilot model mapping to omit non-chat `/models` entries and enable image input for models whose capabilities indicate vision support
+
+### Fixed
+
+- Fixed long-context variant pricing to use `billing.token_prices.long_context` rates instead of default model pricing
+- Fixed `mapModel` handling in OpenAI-compatible discovery so returning `null` now skips a model entry rather than falling back to defaults
+- Fixed model ID precedence so a real upstream Copilot model id is kept when it conflicts with a synthesized `-1m` variant
+
+## [15.11.1] - 2026-06-11
+
+### Fixed
+
+- Fixed NVIDIA NIM Qwen turns failing with `400 Validation: Unsupported parameter(s): enable_thinking`. NIM's chat-completions schema is `additionalProperties: false` and exposes thinking via the vLLM convention `chat_template_kwargs.enable_thinking`; `buildOpenAICompat` was sending top-level `enable_thinking` for every `qwen/*` id regardless of host. Registered `nvidia` as a known host (`integrate.api.nvidia.com`) and routed NVIDIA-hosted Qwen models to `thinkingFormat: "qwen-chat-template"` ([#2299](https://github.com/can1357/oh-my-pi/issues/2299)).
+- Fixed Moonshot/Kimi native OpenAI-compatible request metadata so Kimi K2 uses `max_tokens` and omits OpenAI-only `store`, restoring first-turn output with `MOONSHOT_API_KEY` ([#2289](https://github.com/can1357/oh-my-pi/issues/2289)).
+
+## [15.11.0] - 2026-06-10
+
+### Fixed
+
+- Fixed `buildModel` so malformed explicit thinking metadata without `efforts` is treated as sparse input and inferred instead of crashing during model resolution ([#2251](https://github.com/can1357/oh-my-pi/issues/2251)).
+
+## [15.10.12] - 2026-06-10
+
+### Added
+
 - Added `grok-composer-2.5-fast` (Cursor "Composer 2.5 Fast") to the xAI Grok OAuth (SuperGrok) catalog: non-reasoning, text-only, 200K context.
+
+### Changed
+
+- Set every xAI Grok OAuth (SuperGrok) curated model's max output tokens to mirror its context window (`grok-build`, `grok-4.3`, `grok-4.20-0309-{reasoning,non-reasoning}`, `grok-4.20-multi-agent-0309`, `grok-composer-2.5-fast`), replacing the `8888` `UNK_MAX_TOKENS` placeholder (and a stale `30000` on three grok-4.x entries). xAI's OAuth `/v1/models` reports no per-request output limit, so the curated catalog now owns `maxTokens` like `contextWindow`, deterministic on both the static-seed and online-overlay paths; the `openai-responses` wire still clamps the actual request to `OPENAI_MAX_OUTPUT_TOKENS` (64k).
+
+### Fixed
+
+- Excluded zero-cost `xai-oauth` subscription entries from the model reference indexes (`buildModelReferenceIndex`, `createReferenceResolver`), so their zero pricing and context-window-sized `maxTokens` cannot outrank paid/public Grok references when resolving custom-provider model identities.
+
+## [15.10.11] - 2026-06-10
+
+### Added
+
 - Added `hostMatchesUrl`, `modelMatchesHost`, and endpoint-shape helpers in the new `hosts` module for consistent provider/baseUrl matching
 - `buildModel(spec)` (`build.ts`) is now the single Model constructor: it materializes the fully-resolved compat record and canonical thinking metadata exactly once (compat first, thinking derived from identity + resolved compat), so `Model.compat` is a required, complete `CompatOf<TApi>` (`ResolvedOpenAICompat`/`ResolvedOpenAIResponsesCompat`/`ResolvedAnthropicCompat`) and request-path code reads fields with zero URL parsing and zero per-request allocation. Sparse user/config overrides live on the new `ModelSpec<TApi>` input shape and survive on `Model.compatConfig` for introspection.
 - Added `ResolvedAnthropicCompat.supportsSamplingParams` (Opus 4.7+/Fable/Mythos reject `temperature`/`top_p`/`top_k` with a 400), baked at build time from model identity so the request path stops re-parsing model ids.
@@ -27,21 +168,6 @@
 
 ### Changed
 
-- Changed catalog metadata to update a model’s per-token pricing to input 0.09 and output 0.18
-- Changed the same cataloged model’s maximum token limit from 384000 to 65536
-- Pinned zai `glm-5.2` to 1M context during catalog generation so endpoint discovery and older fallbacks cannot regress it to 200k.
-- Replaced the hand-maintained `zhipu-coding-plan` GLM reasoning allowlist and vision regex with a `parseGlmModel` family classifier in `identity/classify.ts` (variant + vision + version), surfaced as `isReasoningGlmModelId` / `isGlmVisionModelId`. Discovery now derives reasoning/vision capability from the GLM family instead of a per-id list, so newly-bumped integers (`glm-5.3`, `glm-6`, …) are covered automatically while `-flash`/`-preview` and the vision `…v` shape stay correctly classified.
-- Model `contextWindow`/`maxTokens` are now `number | null`; discovery emits `null` when a provider reports no limit, replacing the `222222`/`8888` (`UNK_CONTEXT_WINDOW`/`UNK_MAX_TOKENS`) sentinels (now removed). Bundled `models.json` unknown limits are `null`.
-- Changed the `github-copilot` model context window to `524288` tokens
-- Changed Fireworks model discovery to source the control-plane `List Models` API (`GET /v1/accounts/fireworks/models?filter=supports_serverless=true`) instead of the OpenAI-compatible `/v1/models` inference listing. The inference endpoint returns a sparse, account-specific subset that omits on-demand serverless models (e.g. `kimi-k2.7-code`), so newly published serverless models stayed invisible in the picker until hand-added to the bundled catalog. The control-plane catalog enumerates every serverless model with capability metadata (`supportsServerless`/`supportsTools`/`supportsImageInput`/`contextLength`/`displayName`), paginated and filtered to tool-capable `READY` entries, then merged with bundled/models.dev references — the Kimi K2 max-output clamp and DeepSeek V4 thinking-toggle strip are preserved, and unbundled models default to reasoning so `buildModel` derives the Fireworks effort map. New serverless releases now surface automatically with no catalog edits.
-- Changed model display names to drop model-extrinsic decorations: gateway author prefixes (`OpenAI: …`, `Google: …`), `(latest)` alias markers, `(Antigravity)` provider attribution, price tiers (`($$$$)`), and promo/lifecycle tags (`(20% off)`, `(retires …)`). `cleanModelName` is applied in `buildModel` (covers live discovery and stale caches) and as a catalog-generator pass; Antigravity discovery no longer appends `(Antigravity)` to display names. Variant tags that map to distinct wire ids (`(Thinking)`, `(free)`, `(Fast)`, dates, regions) are preserved.
-- Changed the `google-antigravity` default model from `gemini-3-pro-high` to `gemini-3.1-pro`
-- Changed `gemini-2.5-flash-thinking` handling from discovery-denylist to collapsing into `gemini-2.5-flash` (thinking-enabled requests route to the `-thinking` backing id)
-- Bumped the model cache schema to v5 so rows predating effort-tier variant collapsing (raw `-low`/`-high`/`-thinking` member ids) are invalidated
-- Changed GitHub Copilot discovery to request `X-GitHub-Api-Version: 2026-06-01` from `api.githubcopilot.com`
-- Changed GitHub Copilot discovery to cap base model `contextWindow` to the default token tier and keep long-context access as the separate `-1m` model entry
-- Changed Copilot model mapping to omit non-chat `/models` entries and enable image input for models whose capabilities indicate vision support
-- Set every xAI Grok OAuth (SuperGrok) curated model's max output tokens to mirror its context window (`grok-build`, `grok-4.3`, `grok-4.20-0309-{reasoning,non-reasoning}`, `grok-4.20-multi-agent-0309`, `grok-composer-2.5-fast`), replacing the `8888` `UNK_MAX_TOKENS` placeholder (and a stale `30000` on three grok-4.x entries). xAI's OAuth `/v1/models` reports no per-request output limit, so the curated catalog now owns `maxTokens` like `contextWindow`, deterministic on both the static-seed and online-overlay paths; the `openai-responses` wire still clamps the actual request to `OPENAI_MAX_OUTPUT_TOKENS` (64k).
 - Changed OpenAI compatibility detection to use shared host classifiers (`modelMatchesHost`/`hostMatchesUrl`) with normalized matching instead of raw URL substring checks
 - Changed `hostMatchesUrl`/`modelMatchesHost` usage in compatibility detection to reduce mismatches across case variants and provider alias hosts
 - Provider catalog entries now carry the runtime API-key env fallback as an ordered `envVars` list; `catalogDiscovery.envVars` became an optional generation-time override (only `cursor` and `vercel-ai-gateway` differ) and `PROVIDER_DESCRIPTORS` materializes the resolved list for `generate-models.ts`.
@@ -52,25 +178,6 @@
 
 ### Fixed
 
-- Fixed MiniMax-M3 catalog context for `minimax` and `minimax-cn` to report the documented 1M long-context tier instead of the upstream 512K pricing boundary ([#2576](https://github.com/can1357/oh-my-pi/issues/2576)).
-- Fixed OpenCode Go MiMo catalog metadata so title generation and other tool-enabled calls omit unsupported `tool_choice` instead of triggering provider 400s ([#2509](https://github.com/can1357/oh-my-pi/issues/2509)).
-- Fixed OpenCode Go `kimi-k2.7-code` catalog metadata so resolve-gate requests use automatic tool selection instead of Moonshot-rejected forced `tool_choice` ([#2546](https://github.com/can1357/oh-my-pi/issues/2546)).
-- Fixed Anthropic compat for the `github-copilot` host so `supportsEagerToolInputStreaming` defaults to `false` there, matching the Copilot proxy which rejects the per-tool `eager_input_streaming` field ([#2558](https://github.com/can1357/oh-my-pi/issues/2558)).
-- Scoped vLLM model cache validity to the discovery base URL so changed endpoints refetch immediately, and bounded built-in vLLM discovery requests with a timeout.
-- Filled missing `contextWindow` and `maxTokens` in generated `models.json` for proxy/reseller variants by inheriting limits from canonical-family and segment-reference models
-- Ignored zero-cost `x-ai` subscription entries as reference sources when backfilling limits so inflated values are not propagated
-- Fixed the model cache opening with `PRAGMA journal_mode=WAL` before `PRAGMA busy_timeout`, so concurrent omp startups could crash inside `getDb()` on `SQLITE_BUSY` during WAL recovery instead of waiting through the transient lock. The busy handler is now installed before the first lock-taking statement ([#2421](https://github.com/can1357/oh-my-pi/issues/2421)).
-- Fixed Antigravity `gemini-3.1-pro --thinking high` failing with `Cloud Code Assist API error (400): Request contains an invalid argument.` — the upstream `gemini-3.1-pro-high` deployment rejects every `streamGenerateContent` request on both CCA endpoints while discovery still advertises it. High effort now routes to `gemini-pro-agent` (the same "Gemini 3.1 Pro (High)" model, verified accepting the identical request body), and the model-cache fingerprint version was bumped (`merge-v2` → `merge-v3`) so existing fresh caches refetch discovery and pick up the corrected routing immediately.
-- Fixed catalog generation to apply effort-tier variant collapsing before provider grouping to ensure collapsed model families are consistently materialized without being impacted by in-loop mutation
-- Fixed Kimi K2.6 OpenAI-compatible compat metadata to use a 300s stream watchdog floor, covering Fire Pass router ids as well as public `kimi-k2.6` ids so long reasoning starts do not hit the generic first-event timeout ([#2366](https://github.com/can1357/oh-my-pi/issues/2366)).
-- Fixed MiniMax M2-family and OpenAI gpt-oss model metadata so OpenAI-compatible catalog entries declare only `low|medium|high` thinking efforts. Their upstreams reject `minimal`, `xhigh`, and Fireworks' `minimal → none` wire mapping, so `fireworks/minimax-m2.7` as the smol auto-thinking classifier model 400ed on every turn. OpenAI-compatible provider effort maps (`Groq qwen/qwen3-32b`, DeepSeek-family, OpenRouter Anthropic adaptive, Fireworks `minimal → none`) now bake into `thinking.effortMap` in catalog metadata instead of `buildOpenAICompat`, and request builders read that field directly. Regenerated `models.json` now makes `disableReasoning` choose `low` for those families while leaving GLM-5.x and other Fireworks models on the existing `minimal → none` path ([#2315](https://github.com/can1357/oh-my-pi/issues/2315)).
-- Fixed long-context variant pricing to use `billing.token_prices.long_context` rates instead of default model pricing
-- Fixed `mapModel` handling in OpenAI-compatible discovery so returning `null` now skips a model entry rather than falling back to defaults
-- Fixed model ID precedence so a real upstream Copilot model id is kept when it conflicts with a synthesized `-1m` variant
-- Fixed NVIDIA NIM Qwen turns failing with `400 Validation: Unsupported parameter(s): enable_thinking`. NIM's chat-completions schema is `additionalProperties: false` and exposes thinking via the vLLM convention `chat_template_kwargs.enable_thinking`; `buildOpenAICompat` was sending top-level `enable_thinking` for every `qwen/*` id regardless of host. Registered `nvidia` as a known host (`integrate.api.nvidia.com`) and routed NVIDIA-hosted Qwen models to `thinkingFormat: "qwen-chat-template"` ([#2299](https://github.com/can1357/oh-my-pi/issues/2299)).
-- Fixed Moonshot/Kimi native OpenAI-compatible request metadata so Kimi K2 uses `max_tokens` and omits OpenAI-only `store`, restoring first-turn output with `MOONSHOT_API_KEY` ([#2289](https://github.com/can1357/oh-my-pi/issues/2289)).
-- Fixed `buildModel` so malformed explicit thinking metadata without `efforts` is treated as sparse input and inferred instead of crashing during model resolution ([#2251](https://github.com/can1357/oh-my-pi/issues/2251)).
-- Excluded zero-cost `xai-oauth` subscription entries from the model reference indexes (`buildModelReferenceIndex`, `createReferenceResolver`), so their zero pricing and context-window-sized `maxTokens` cannot outrank paid/public Grok references when resolving custom-provider model identities.
 - Fixed Anthropic official-endpoint detection to require strict HTTPS hostname matching so non-official or lookalike URLs are no longer treated as official Anthropic hosts
 - Fixed Ollama Cloud dynamic discovery so same-id matches from other providers no longer supply context-window or max-output-token limits for discovered models.
 - Wired `@oh-my-pi/pi-catalog` into the release publish package list, tarball install smoke test, and root `bun generate-models` script.
@@ -80,25 +187,3 @@
 ### Removed
 
 - Removed the runtime enrichment layer: `enrichModelThinking` (and its non-enumerable memo-slot cache), `refreshModelThinking`, `modelOmitsReasoningEffort`, and the `model-thinking` re-exports of generator-only policies. Thinking metadata is resolved exactly once inside `buildModel`; runtime helpers (`getSupportedEfforts`, `clampThinkingLevelForModel`, `requireSupportedEffort`, the effort mappers) are pure field reads.
-
-## [15.13.0] - 2026-06-14
-
-## [15.12.6] - 2026-06-14
-
-## [15.12.4] - 2026-06-13
-
-## [15.11.8] - 2026-06-12
-
-## [15.11.7] - 2026-06-12
-
-## [15.11.4] - 2026-06-12
-
-## [15.11.3] - 2026-06-11
-
-## [15.11.1] - 2026-06-11
-
-## [15.11.0] - 2026-06-10
-
-## [15.10.12] - 2026-06-10
-
-## [15.10.11] - 2026-06-10

package/dist/types/identity/family.d.ts

@@ -37,6 +37,8 @@ export declare function isMinimaxM2FamilyModelId(modelId: string): boolean;
  * and `none`.
  */
 export declare function isOpenAIGptOssModelId(modelId: string): boolean;
+/** OpenAI model ids (gpt-*, o1-*, o3-*, o4-*, or prefixed with openai/). */
+export declare function isOpenAIModelId(modelId: string): boolean;
 /**
  * Reasoning-capable GLM coding SKUs: glm-4.5 and up on the base / `-air` /
  * `-turbo` lines. Excludes the vision (`…v`) shape, the non-reasoning

package/dist/types/identity/index.d.ts

@@ -7,3 +7,4 @@ export * from "./markers";
 export * from "./priority";
 export * from "./reference";
 export * from "./selection";
+export * from "./tool-syntax";

package/dist/types/identity/tool-syntax.d.ts

@@ -0,0 +1,3 @@
+export type ToolCallSyntax = "glm" | "hermes" | "kimi" | "xml" | "anthropic" | "deepseek" | "harmony" | "pi" | "qwen3";
+export declare const FALLBACK_TOOL_SYNTAX: ToolCallSyntax;
+export declare function preferredToolSyntax(modelId: string): ToolCallSyntax;

package/dist/types/types.d.ts

@@ -371,6 +371,13 @@ export interface Model<TApi extends Api = Api> {
     baseUrl: string;
     reasoning: boolean;
     input: ("text" | "image")[];
+    /**
+     * Native provider tool-call support. `false` is the only unsupported signal:
+     * `true` and `undefined` both mean callers may use native tools. Catalog and
+     * discovery sources should set this sparsely when an upstream explicitly
+     * reports that native tool calling is unsupported.
+     */
+    supportsTools?: boolean;
     cost: {
         input: number;
         output: number;

package/package.json

@@ -1,7 +1,7 @@
 {
 	"type": "module",
 	"name": "@oh-my-pi/pi-catalog",
-	"version": "15.13.0",
+	"version": "15.13.2",
 	"description": "Model catalog for omp: bundled model database, provider discovery descriptors, model identity, classification, and equivalence",
 	"homepage": "https://omp.sh",
 	"author": "Can Boluk",
@@ -34,11 +34,11 @@
 	},
 	"dependencies": {
 		"@bufbuild/protobuf": "^2.12.0",
-		"@oh-my-pi/pi-utils": "15.13.0",
+		"@oh-my-pi/pi-utils": "15.13.2",
 		"zod": "^4"
 	},
 	"devDependencies": {
-		"@oh-my-pi/pi-ai": "15.13.0",
+		"@oh-my-pi/pi-ai": "15.13.2",
 		"@types/bun": "^1.3.14"
 	},
 	"engines": {

package/src/identity/family.ts

@@ -78,6 +78,11 @@ export function isOpenAIGptOssModelId(modelId: string): boolean {
 	return /(^|\/)gpt-oss[-:]/i.test(modelId);
 }
 
+/** OpenAI model ids (gpt-*, o1-*, o3-*, o4-*, or prefixed with openai/). */
+export function isOpenAIModelId(modelId: string): boolean {
+	return /(^|\/)(gpt|o1|o3|o4)[-.]/i.test(modelId) || modelId.toLowerCase().includes("openai/");
+}
+
 /**
  * Reasoning-capable GLM coding SKUs: glm-4.5 and up on the base / `-air` /
  * `-turbo` lines. Excludes the vision (`…v`) shape, the non-reasoning
@@ -114,6 +119,8 @@ export function isGlmVisionModelId(modelId: string): boolean {
 export function modelFamilyToken(modelId: string): string {
 	const parsed = parseKnownModel(modelId);
 	if (parsed.family !== "unknown") return parsed.family;
+	if (isClaudeModelId(modelId) || isAnthropicNamespacedModelId(modelId)) return "anthropic";
+	if (isOpenAIModelId(modelId)) return "openai";
 	if (isKimiModelId(modelId)) return "kimi";
 	if (isQwenModelId(modelId)) return "qwen";
 	if (isMinimaxM2FamilyModelId(modelId)) return "minimax";

package/src/identity/index.ts

@@ -7,3 +7,4 @@ export * from "./markers";
 export * from "./priority";
 export * from "./reference";
 export * from "./selection";
+export * from "./tool-syntax";

package/src/identity/tool-syntax.ts

@@ -0,0 +1,25 @@
+import { modelFamilyToken } from "./family";
+
+export type ToolCallSyntax = "glm" | "hermes" | "kimi" | "xml" | "anthropic" | "deepseek" | "harmony" | "pi" | "qwen3";
+
+export const FALLBACK_TOOL_SYNTAX: ToolCallSyntax = "xml";
+
+export function preferredToolSyntax(modelId: string): ToolCallSyntax {
+	switch (modelFamilyToken(modelId)) {
+		case "anthropic":
+			return "anthropic";
+		case "glm":
+			return "glm";
+		case "kimi":
+			return "kimi";
+		case "qwen":
+			return "qwen3";
+		case "deepseek":
+			return "deepseek";
+		case "openai":
+		case "gpt-oss":
+			return "harmony";
+		default:
+			return FALLBACK_TOOL_SYNTAX;
+	}
+}

package/src/models.json

@@ -13692,6 +13692,39 @@
 				"requiresAssistantContentForToolCalls": true
 			}
 		},
+		"flux-1-schnell-fp8": {
+			"id": "flux-1-schnell-fp8",
+			"name": "FLUX.1 [schnell] FP8",
+			"api": "openai-completions",
+			"provider": "fireworks",
+			"baseUrl": "https://api.fireworks.ai/inference/v1",
+			"reasoning": true,
+			"input": [
+				"text"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": null,
+			"maxTokens": null,
+			"supportsTools": false,
+			"thinking": {
+				"mode": "effort",
+				"efforts": [
+					"minimal",
+					"low",
+					"medium",
+					"high",
+					"xhigh"
+				],
+				"effortMap": {
+					"minimal": "none"
+				}
+			}
+		},
 		"glm-5": {
 			"id": "glm-5",
 			"name": "GLM-5",
@@ -13783,6 +13816,26 @@
 				]
 			}
 		},
+		"gpt-oss-20b": {
+			"id": "gpt-oss-20b",
+			"name": "OpenAI gpt-oss-20b",
+			"api": "openai-completions",
+			"provider": "fireworks",
+			"baseUrl": "https://api.fireworks.ai/inference/v1",
+			"reasoning": false,
+			"input": [
+				"text"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": 131072,
+			"maxTokens": 65536,
+			"supportsTools": false
+		},
 		"kimi-k2.5": {
 			"id": "kimi-k2.5",
 			"name": "Kimi K2.5",
@@ -14003,6 +14056,70 @@
 				}
 			}
 		},
+		"qwen3-embedding-8b": {
+			"id": "qwen3-embedding-8b",
+			"name": "Qwen3 Embedding 8B",
+			"api": "openai-completions",
+			"provider": "fireworks",
+			"baseUrl": "https://api.fireworks.ai/inference/v1",
+			"reasoning": true,
+			"input": [
+				"text"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": 40960,
+			"maxTokens": null,
+			"supportsTools": false,
+			"thinking": {
+				"mode": "effort",
+				"efforts": [
+					"minimal",
+					"low",
+					"medium",
+					"high"
+				],
+				"effortMap": {
+					"minimal": "none"
+				}
+			}
+		},
+		"qwen3-reranker-8b": {
+			"id": "qwen3-reranker-8b",
+			"name": "Qwen3 Reranker 8B",
+			"api": "openai-completions",
+			"provider": "fireworks",
+			"baseUrl": "https://api.fireworks.ai/inference/v1",
+			"reasoning": true,
+			"input": [
+				"text"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": 40960,
+			"maxTokens": null,
+			"supportsTools": false,
+			"thinking": {
+				"mode": "effort",
+				"efforts": [
+					"minimal",
+					"low",
+					"medium",
+					"high"
+				],
+				"effortMap": {
+					"minimal": "none"
+				}
+			}
+		},
 		"qwen3.6-plus": {
 			"id": "qwen3.6-plus",
 			"name": "Qwen3.6 Plus",
@@ -16509,6 +16626,64 @@
 					"high"
 				]
 			}
+		},
+		"gemma-4-E2B-it": {
+			"id": "gemma-4-E2B-it",
+			"name": "Gemma 4 E2B IT",
+			"api": "google-generative-ai",
+			"provider": "google",
+			"baseUrl": "https://generativelanguage.googleapis.com/v1beta",
+			"reasoning": true,
+			"input": [
+				"text",
+				"image"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": 131072,
+			"maxTokens": 8192,
+			"thinking": {
+				"mode": "budget",
+				"efforts": [
+					"minimal",
+					"low",
+					"medium",
+					"high"
+				]
+			}
+		},
+		"gemma-4-E4B-it": {
+			"id": "gemma-4-E4B-it",
+			"name": "Gemma 4 E4B IT",
+			"api": "google-generative-ai",
+			"provider": "google",
+			"baseUrl": "https://generativelanguage.googleapis.com/v1beta",
+			"reasoning": true,
+			"input": [
+				"text",
+				"image"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": 131072,
+			"maxTokens": 8192,
+			"thinking": {
+				"mode": "budget",
+				"efforts": [
+					"minimal",
+					"low",
+					"medium",
+					"high"
+				]
+			}
 		}
 	},
 	"google-antigravity": {
@@ -75419,6 +75594,34 @@
 				"requiresEffort": true
 			}
 		},
+		"qwen/qwen3-asr-flash": {
+			"id": "qwen/qwen3-asr-flash",
+			"name": "Qwen3-ASR-Flash",
+			"api": "openai-completions",
+			"provider": "zenmux",
+			"baseUrl": "https://zenmux.ai/api/v1",
+			"reasoning": true,
+			"input": [
+				"text"
+			],
+			"cost": {
+				"input": 0,
+				"output": 0,
+				"cacheRead": 0,
+				"cacheWrite": 0
+			},
+			"contextWindow": 1000000,
+			"maxTokens": null,
+			"thinking": {
+				"mode": "effort",
+				"efforts": [
+					"minimal",
+					"low",
+					"medium",
+					"high"
+				]
+			}
+		},
 		"qwen/qwen3-coder": {
 			"id": "qwen/qwen3-coder",
 			"name": "Qwen3-Coder",

package/src/provider-models/openai-compat.ts

@@ -1160,6 +1160,7 @@ function mapFireworksControlPlaneModel(
 ): ModelSpec<"openai-completions"> {
 	const name = toModelName(record.displayName, reference?.name ?? publicModelId);
 	const supportsImage = toBoolean(record.supportsImageInput) === true;
+	const supportsTools = toBoolean(record.supportsTools);
 	const contextWindow = toPositiveNumber(record.contextLength, reference?.contextWindow ?? null);
 	// The control plane reports no max-output budget; default the Kimi family to
 	// its published cap, everyone else to the discovery fallback, then clamp.
@@ -1192,6 +1193,7 @@ function mapFireworksControlPlaneModel(
 		input: supportsImage ? ["text", "image"] : (reference?.input ?? ["text"]),
 		contextWindow,
 		maxTokens,
+		...(supportsTools === false ? { supportsTools: false } : {}),
 	};
 	return stripFireworksDeepSeekThinkingToggle(model, publicModelId);
 }
@@ -1240,7 +1242,6 @@ async function fetchFireworksServerlessModels(options: {
 			if (!isRecord(entry)) continue;
 			const record = entry as FireworksControlPlaneModel;
 			if (toBoolean(record.supportsServerless) !== true) continue;
-			if (toBoolean(record.supportsTools) !== true) continue;
 			if (typeof record.state === "string" && record.state !== "READY") continue;
 			const wireName = typeof record.name === "string" ? record.name : "";
 			if (!wireName) continue;
@@ -1396,6 +1397,7 @@ function mapWaferModel(
 	const capabilities = wafer?.capabilities ?? {};
 	const reasoning = capabilities.reasoning === true;
 	const vision = capabilities.vision === true;
+	const supportsTools = toBoolean(capabilities.tools) === false ? false : undefined;
 	const contextWindow = toPositiveNumber(
 		wafer?.context_length,
 		toPositiveNumber((entry as { max_model_len?: unknown }).max_model_len, defaults.contextWindow),
@@ -1434,6 +1436,7 @@ function mapWaferModel(
 		cost,
 		contextWindow,
 		maxTokens,
+		...(supportsTools === false ? { supportsTools } : {}),
 	};
 	if (reasoning) {
 		// Wafer's `wafer.provider` envelope tells us which upstream backend serves
@@ -2928,6 +2931,7 @@ export function mapModelsDevToModels(
 				},
 				contextWindow: toPositiveNumber(m.limit?.context, desc.defaultContextWindow ?? null),
 				maxTokens: toPositiveNumber(m.limit?.output, desc.defaultMaxTokens ?? null),
+				...(m.tool_call === false ? { supportsTools: false } : {}),
 				...(desc.compat && { compat: desc.compat }),
 				...(desc.headers && { headers: { ...desc.headers } }),
 			};

package/src/types.ts

@@ -436,6 +436,13 @@ export interface Model<TApi extends Api = Api> {
 	baseUrl: string;
 	reasoning: boolean;
 	input: ("text" | "image")[];
+	/**
+	 * Native provider tool-call support. `false` is the only unsupported signal:
+	 * `true` and `undefined` both mean callers may use native tools. Catalog and
+	 * discovery sources should set this sparsely when an upstream explicitly
+	 * reports that native tool calling is unsupported.
+	 */
+	supportsTools?: boolean;
 	cost: {
 		input: number; // $/million tokens
 		output: number; // $/million tokens