@oh-my-pi/pi-catalog 18.14.2 → 18.15.0

package/CHANGELOG.md

@@ -2,6 +2,17 @@
 
 ## [Unreleased]
 
+## [18.15.0] - 2026-11-01
+
+### Added
+
+- Added Modular Cloud (`modular`) as a provider: `/login modular` with an API key, models and prices from the models.dev `modular` entry, and per-model reasoning controls for MiniMax M3.1 Flash, MiniMax M3, and Gemma 4 31B ([#17334](https://github.com/can1357/oh-my-pi/pull/17334) by [@tommyjtl](https://github.com/tommyjtl))
+- SambaNova Cloud follow-up turns now reuse the provider's prefix cache, so multi-turn conversations stop re-processing earlier context ([#17555](https://github.com/can1357/oh-my-pi/pull/17555) by [@rjuez00](https://github.com/rjuez00))
+
+### Fixed
+
+- Fixed newly discovered OpenAI-compatible models ignoring provider-reported context and output token limits ([#17561](https://github.com/can1357/oh-my-pi/pull/17561) by [@wolfiesch](https://github.com/wolfiesch))
+
 ## [18.14.2] - 2026-10-31
 
 ### Breaking Changes

package/dist/types/compat/auth-ids.d.ts

@@ -1,4 +1,4 @@
 /** Every provider with an `auth/<id>.kdl` policy. */
-export type AuthProviderId = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";
+export type AuthProviderId = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";
 /** Providers whose policy declares a `login` flow (the `/login` roster). */
-export type LoginProviderId = "abliteration" | "aiand" | "alibaba-coding-plan" | "alibaba-token-plan" | "anthropic" | "baseten" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google-antigravity" | "google-gemini-cli" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "meta" | "minimax-code" | "minimax-code-cn" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";
+export type LoginProviderId = "abliteration" | "aiand" | "alibaba-coding-plan" | "alibaba-token-plan" | "anthropic" | "baseten" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "exa" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google-antigravity" | "google-gemini-cli" | "helmcode" | "huggingface" | "inception" | "kagi" | "kilo" | "kimi-code" | "litellm" | "llama.cpp" | "lm-studio" | "meta" | "minimax-code" | "minimax-code-cn" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai-codex" | "openai-codex-device" | "opencode-go" | "opencode-zen" | "openrouter" | "parallel" | "perplexity" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stencil" | "stepfun" | "synthetic" | "tavily" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zai-coding-plan" | "zenmux" | "zhipu-coding-plan";

package/dist/types/compat/provider-ids.d.ts

@@ -1,2 +1,2 @@
 /** Every chat-model provider with a catalog entry (`default-model` in its provider rules). */
-export type KnownProvider = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kilo" | "kimi-code" | "litellm" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "opencode-go" | "opencode-zen" | "openrouter" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stepfun" | "synthetic" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zenmux" | "zhipu-coding-plan";
+export type KnownProvider = "abliteration" | "aiand" | "aimlapi" | "alibaba-coding-plan" | "alibaba-token-plan" | "amazon-bedrock" | "anthropic" | "apple" | "azure" | "baseten" | "bedrock-mantle" | "cerebras" | "charm-hyper" | "cline-pass" | "cloudflare-ai-gateway" | "cloudflare-workers-ai" | "commandcode" | "coralbricks" | "coreweave" | "cursor" | "deepinfra" | "deepseek" | "devin" | "factory-droid" | "firepass" | "fireworks" | "github-copilot" | "gitlab-duo" | "gitlab-duo-agent" | "gmi-cloud" | "google" | "google-antigravity" | "google-gemini-cli" | "google-vertex" | "groq" | "helmcode" | "huggingface" | "inception" | "kilo" | "kimi-code" | "litellm" | "lm-studio" | "local" | "meta" | "minimax" | "minimax-code" | "minimax-code-cn" | "mistral" | "modular" | "moonshot" | "muse-code" | "nanogpt" | "novita" | "nvidia" | "ollama" | "ollama-cloud" | "openai" | "openai-codex" | "opencode-go" | "opencode-zen" | "openrouter" | "qianfan" | "qwen-portal" | "sakana" | "sambanova" | "siliconflow" | "siliconflow-cn" | "singularityapi-dev" | "singularityapi-tech" | "snowflake" | "stepfun" | "synthetic" | "together" | "typesafe" | "umans" | "venice" | "vercel-ai-gateway" | "vllm" | "wafer-serverless" | "web" | "xai" | "xai-oauth" | "xiaomi" | "xiaomi-token-plan-ams" | "xiaomi-token-plan-cn" | "xiaomi-token-plan-sgp" | "yolo-auto" | "zai" | "zenmux" | "zhipu-coding-plan";

package/dist/types/types.d.ts

@@ -1467,8 +1467,9 @@ export interface Model<TApi extends Api = Api> {
     useResponsesLite?: boolean;
     /**
      * Service-tier ids the provider advertises for this model (Codex discovery
-     * `service_tiers[].id`, e.g. `priority`, `ultrafast`). Absent when the
-     * provider publishes no per-model tier list.
+     * `service_tiers[].id`, or an OpenAI-compatible `/v1/models` row's
+     * `service_tiers`, e.g. `priority`, `ultrafast`). Absent when the provider
+     * publishes no per-model tier list.
      */
     serviceTiers?: readonly string[];
     /** Preferred model to switch to when context promotion is triggered (model id or provider/id). */

package/dist/types/utils.d.ts

@@ -11,6 +11,14 @@ export declare function toPositiveNumber(value: unknown, fallback: number | null
 /** Positive finite number, or `null` when the value is missing/non-positive. */
 export declare function toPositiveNumberOrNull(value: unknown): number | null;
 export declare function toBoolean(value: unknown): boolean | undefined;
+/**
+ * Service-tier ids from a discovery row's `service_tiers` array, deduplicated
+ * in advertisement order. Entries may be codex-rs `{ id, name, description }`
+ * objects or bare id strings (OpenAI-compatible relays). Returns `undefined`
+ * when the value is not an array; an empty array is kept, since it is a
+ * statement that the model offers no optional tier.
+ */
+export declare function parseServiceTierIds(value: unknown): string[] | undefined;
 export declare function isAnthropicOAuthToken(key: string): boolean;
 /**
  * Normalize a model display name: drop the gateway author prefix and

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@oh-my-pi/pi-catalog",
-	"version": "18.14.2",
+	"version": "18.15.0",
 	"description": "Model catalog for omp: bundled model database, provider discovery descriptors, model identity, classification, and equivalence",
 	"keywords": [
 		"ai",
@@ -88,12 +88,12 @@
 		"gen:proto": "bun scripts/generate-protocols.ts"
 	},
 	"dependencies": {
-		"@oh-my-pi/omptype": "18.14.2",
-		"@oh-my-pi/pi-utils": "18.14.2"
+		"@oh-my-pi/omptype": "18.15.0",
+		"@oh-my-pi/pi-utils": "18.15.0"
 	},
 	"devDependencies": {
 		"@bgotink/kdl": "0.4.0",
-		"@oh-my-pi/pi-ai": "18.14.2",
+		"@oh-my-pi/pi-ai": "18.15.0",
 		"@types/bun": "^1.3.14"
 	},
 	"engines": {

package/src/compat/auth-ids.ts

@@ -53,6 +53,7 @@ export type AuthProviderId =
 	| "minimax-code"
 	| "minimax-code-cn"
 	| "mistral"
+	| "modular"
 	| "moonshot"
 	| "muse-code"
 	| "nanogpt"
@@ -143,6 +144,7 @@ export type LoginProviderId =
 	| "meta"
 	| "minimax-code"
 	| "minimax-code-cn"
+	| "modular"
 	| "moonshot"
 	| "muse-code"
 	| "nanogpt"

package/src/compat/provider-ids.ts

@@ -50,6 +50,7 @@ export type KnownProvider =
 	| "minimax-code"
 	| "minimax-code-cn"
 	| "mistral"
+	| "modular"
 	| "moonshot"
 	| "muse-code"
 	| "nanogpt"

package/src/compat/rules/auth/_order.kdl

@@ -10,4 +10,4 @@ login-order "openai-codex" "anthropic" "zai" "zai-coding-plan" "kimi-code" "open
 	"qianfan" "venice" "siliconflow" "siliconflow-cn" "synthetic" "nanogpt" "wafer-serverless" "coreweave" \
 	"vercel-ai-gateway" "cloudflare-ai-gateway" "cloudflare-workers-ai" "snowflake" "litellm" "kilo" "zenmux" "opencode-zen" "opencode-go" \
 	"yolo-auto" "tavily" "kagi" "exa" "parallel" "typesafe" "ollama" "ollama-cloud" "lm-studio" "llama.cpp" "vllm" \
-	"gmi-cloud" "stencil" "singularityapi-dev" "singularityapi-tech" "coralbricks"
+	"gmi-cloud" "stencil" "singularityapi-dev" "singularityapi-tech" "coralbricks" "modular"

package/src/compat/rules/auth/modular.kdl

@@ -0,0 +1,9 @@
+auth "modular" {
+	name "Modular"
+	login "api-key" {
+		auth-url "https://console.modular.com/api_keys"
+		instructions "Create an API key in the Modular Console"
+		prompt "Paste your Modular API key" placeholder="..."
+		validate "models-endpoint" label="Modular" url="https://api.modular.com/v1/models"
+	}
+}

package/src/compat/rules/providers/modular.kdl

@@ -0,0 +1,35 @@
+// Provider-wire compat for "modular". Modular Cloud serves open models on an
+// OpenAI-compatible API. The provider is catalog-only: models, prices, and
+// limits come from the models.dev `modular` entry. Reasoning controls differ
+// per model and were measured against the API.
+
+provider "modular" {
+	default-model "google/gemma-4-31b-it"
+	env "MODULAR_API_KEY"
+
+	thinking-mode "effort"
+	// The API takes OpenRouter's `reasoning` object for every model, so a model
+	// models.dev adds later works without its own block: on sends
+	// `reasoning.effort`, off sends `reasoning.enabled = false`.
+	thinking-format "openrouter"
+
+	// MiniMax M3.1 Flash takes a nested `reasoning.effort` and always thinks:
+	// `reasoning.enabled = false` still returns reasoning tokens.
+	models "MiniMaxAI/MiniMax-M3.1-Flash" {
+		thinking-requires-effort #true
+		thinking-efforts "low" "medium" "high" "xhigh" "max"
+	}
+	// MiniMax M3 and Gemma 4 31B switch reasoning with `reasoning.enabled` and
+	// ignore any effort, so the ladder is one on-rung beside off: off sends
+	// `enabled: false`, and on replaces the effort with `enabled: true`.
+	models "minimax/minimax-m3" "google/gemma-4-31b-it" {
+		thinking-efforts "high"
+		when-thinking {
+			extra-body {
+				reasoning {
+					enabled #true
+				}
+			}
+		}
+	}
+}

package/src/compat/rules/providers/sambanova.kdl

@@ -28,6 +28,8 @@ provider "sambanova" {
 	// (https://docs.sambanova.ai/docs/api-reference/chat-completions/create-chat-based-completion).
 	thinking-mode "effort"
 	thinking-format "openai"
+	// Replay the streamed `reasoning` field so turn-N prompts hit the gateway's prefix cache.
+	replay-reasoning-content #true
 	class "gpt-oss" {
 		thinking-efforts "low" "medium" "high"
 		thinking-upgrade-neutral #true