@diegopetrucci/pi-extensions 0.1.67 → 0.1.69

package/extensions/dynamic-context-pruning/index.ts

@@ -75,6 +75,54 @@ export interface MinimalSessionEntry {
 	[key: string]: unknown;
 }
 
+export interface MinimalProjectedSessionEntry {
+	sourceEntry: MinimalSessionEntry;
+	messages: MinimalMessage[];
+}
+
+export interface MinimalSessionProjection {
+	entries: MinimalProjectedSessionEntry[];
+	messages?: MinimalMessage[];
+}
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+	return typeof value === "object" && value !== null;
+}
+
+function isMinimalMessage(value: unknown): value is MinimalMessage {
+	return isRecord(value) && typeof value.role === "string" && value.role.length > 0;
+}
+
+function isProjectionSourceEntry(value: unknown): value is MinimalSessionEntry {
+	return isRecord(value)
+		&& typeof value.type === "string"
+		&& value.type.length > 0
+		&& typeof value.id === "string"
+		&& value.id.length > 0;
+}
+
+export function normalizeSessionProjection(value: unknown): MinimalSessionProjection | undefined {
+	if (!isRecord(value) || !Array.isArray(value.entries)) return undefined;
+
+	const entries: MinimalProjectedSessionEntry[] = [];
+	for (const candidate of value.entries) {
+		if (!isRecord(candidate) || !isProjectionSourceEntry(candidate.sourceEntry) || !Array.isArray(candidate.messages)) return undefined;
+		if (!candidate.messages.every(isMinimalMessage)) return undefined;
+		entries.push({
+			sourceEntry: candidate.sourceEntry,
+			messages: candidate.messages,
+		});
+	}
+
+	if ("messages" in value && (!Array.isArray(value.messages) || !value.messages.every(isMinimalMessage))) return undefined;
+	const messages = Array.isArray(value.messages) ? value.messages : entries.flatMap((entry) => entry.messages);
+	return { entries, messages };
+}
+
+export function sessionProjectionToMessages(projection: MinimalSessionProjection): MinimalMessage[] {
+	return projection.messages ?? projection.entries.flatMap((entry) => entry.messages);
+}
+
 // ============================================================================
 // Constants
 // ============================================================================
@@ -114,6 +162,22 @@ const DEFAULT_CACHED_PRICE_RATIO = 0.1;
  */
 const DEFAULT_BREAK_EVEN_THRESHOLD = 22;
 
+/**
+ * Default mid_loop break-even threshold (pe-zy4s follow-up). The package's own
+ * representative-corpus benchmark (turn-START state definition — the same one
+ * the runtime `context` handler uses) favored rejecting mid_loop candidates
+ * at cachedPriceRatio r=0.1 (mid_loop-optimal T=1, total realized
+ * net benefit ~0.0). This supports a conservative default for the measured
+ * corpus, not a universal claim that mid_loop pruning cannot pay off.
+ * The reported 2026-09-08 byte-level capture (a mid_loop prune at T=22
+ * invalidating 81% of a warm request prefix) illustrates cache invalidation,
+ * not an inability to amortize it: enough subsequent calls can recoup a
+ * one-time bust. Idle keeps T=22; turn-start classification does not imply
+ * the provider cache is cold. Users who want the old behavior can override
+ * via gate.breakEvenThresholdByState.mid_loop.
+ */
+const DEFAULT_MID_LOOP_BREAK_EVEN_THRESHOLD = 1;
+
 /** Tool names never pruned by default: orchestration/subagent-style tools and todo-like state. */
 const DEFAULT_PROTECTED_TOOL_NAMES = [
 	"todo",
@@ -222,20 +286,19 @@ export type GateMode = "on" | "off" | "always-apply";
  * turn-START "idle" population (calls made at the start of a turn) carries
  * essentially ALL the realized net benefit (idle-optimal T=22, total realized
  * net benefit ~20,594 token-units), while "mid_loop" carries essentially NONE
- * (mid_loop-optimal T=1, total realized net benefit ~0.0 -- i.e. no mid_loop
- * candidate is ever worth accepting at r=0.1 under this definition). This is
+ * (mid_loop-optimal T=1, total realized net benefit ~0.0 on this corpus). This is
  * the OPPOSITE of what the pe-c5n9 evidence (measured under the old,
  * non-runtime-observable turn-END definition) suggested.
  *
  * Per the pe-zy4s decision rule (stricter idle default only if idle itself is
  * shown to be ~worthless): idle is NOT worthless here -- it is the ONLY state
- * carrying value at r=0.1 -- so no state-conditioned strictness is justified
- * in either direction. The per-state split is therefore kept at PARITY
- * (`breakEvenThresholdByState: { idle: DEFAULT_BREAK_EVEN_THRESHOLD, mid_loop:
- * DEFAULT_BREAK_EVEN_THRESHOLD }`), both config-overridable. A future ticket
- * could consider a stricter mid_loop default given this evidence, but that is
- * a new, unreviewed direction outside this ticket's authorized decision rule
- * and is left for a follow-up.
+ * carrying value at r=0.1 -- so the idle default stays at
+ * DEFAULT_BREAK_EVEN_THRESHOLD. The mid_loop default, by the same benchmark,
+ * is now DEFAULT_MID_LOOP_BREAK_EVEN_THRESHOLD (=1): the measured candidates
+ * did not justify a more permissive threshold at r=0.1. Mid_loop pruning can
+ * still amortize over enough subsequent calls; state alone does not establish
+ * cache warmth or profitability. Both defaults remain config-overridable via
+ * gate.breakEvenThresholdByState.
  */
 export type AgentState = "idle" | "mid_loop";
 
@@ -285,7 +348,7 @@ export function defaultConfig(): DynamicContextPruningConfig {
 			mode: "on",
 			cachedPriceRatio: DEFAULT_CACHED_PRICE_RATIO,
 			breakEvenThreshold: DEFAULT_BREAK_EVEN_THRESHOLD,
-			breakEvenThresholdByState: { idle: DEFAULT_BREAK_EVEN_THRESHOLD, mid_loop: DEFAULT_BREAK_EVEN_THRESHOLD },
+			breakEvenThresholdByState: { idle: DEFAULT_BREAK_EVEN_THRESHOLD, mid_loop: DEFAULT_MID_LOOP_BREAK_EVEN_THRESHOLD },
 		},
 	};
 }
@@ -684,13 +747,7 @@ export function findToolCallPairIndices(
 	return pair;
 }
 
-/**
- * Fallback correlation for decisions that reference a session entry rather
- * than a toolCallId (reserved for future non-tool-call strategies). Zips
- * message-producing session entries against the current context projection
- * by position; entry/message counts are expected to match in the common
- * case, but this degrades gracefully otherwise.
- */
+/** Map raw session-entry provenance to projected message positions when Pi exposes it. */
 function expectedRoleForEntry(entry: MinimalSessionEntry): string | undefined {
 	if (entry.type === "message") return entry.message?.role;
 	if (entry.type === "custom_message") return "custom";
@@ -699,15 +756,53 @@ function expectedRoleForEntry(entry: MinimalSessionEntry): string | undefined {
 	return undefined;
 }
 
-export function buildEntryIdToMessageIndexMap(
+function buildProjectedEntryIdToMessageIndexMap(
 	entries: MinimalSessionEntry[],
 	messages: MinimalMessage[],
+	projection: MinimalSessionProjection,
 ): Map<string, number> {
 	const map = new Map<string, number>();
+	const entryIds = new Set(entries.map((entry) => entry.id));
+	let cursor = 0;
+
+	for (const projectedEntry of projection.entries) {
+		let firstMessageIndex: number | undefined;
+		for (const projectedMessage of projectedEntry.messages) {
+			let messageIndex = messages.indexOf(projectedMessage, cursor);
+			if (messageIndex < cursor) messageIndex = -1;
+
+			if (messageIndex < 0 && typeof projectedMessage.timestamp === "number") {
+				messageIndex = messages.findIndex(
+					(candidate, index) =>
+						index >= cursor && candidate.role === projectedMessage.role && candidate.timestamp === projectedMessage.timestamp,
+				);
+			}
+			if (messageIndex < 0) continue;
+			firstMessageIndex ??= messageIndex;
+			cursor = messageIndex + 1;
+		}
+
+		if (firstMessageIndex !== undefined && entryIds.has(projectedEntry.sourceEntry.id)) {
+			map.set(projectedEntry.sourceEntry.id, firstMessageIndex);
+		}
+	}
+	return map;
+}
+
+export function buildEntryIdToMessageIndexMap(
+	entries: MinimalSessionEntry[],
+	messages: MinimalMessage[],
+	projection?: MinimalSessionProjection,
+): Map<string, number> {
+	if (projection) return buildProjectedEntryIdToMessageIndexMap(entries, messages, projection);
+
+	// Older runtimes expose only raw branch entries; preserve their positional
+	// correlation and use role/timestamp only when the counts diverge.
 	const producing = entries.filter((entry) =>
 		entry.type === "message" || entry.type === "custom_message" || entry.type === "branch_summary" || entry.type === "compaction",
 	);
 	const length = Math.min(producing.length, messages.length);
+	const map = new Map<string, number>();
 	for (let i = 0; i < length; i++) {
 		const entry = producing[i];
 		const expectedRole = expectedRoleForEntry(entry);
@@ -732,15 +827,21 @@ export function findMessageIndexByRoleAndTimestamp(
 	return index >= 0 ? index : undefined;
 }
 
-/** Resolve a session entry id to a message index using order-zip, then role+timestamp fallback. */
+/** Resolve a session entry id using projection provenance, then legacy role+timestamp fallback. */
 export function resolveMessageIndexForEntry(
 	entryId: string,
 	entries: MinimalSessionEntry[],
 	messages: MinimalMessage[],
+	projection?: MinimalSessionProjection,
 ): number | undefined {
-	const zipMap = buildEntryIdToMessageIndexMap(entries, messages);
+	const zipMap = buildEntryIdToMessageIndexMap(entries, messages, projection);
 	const zipped = zipMap.get(entryId);
 	if (zipped !== undefined) return zipped;
+
+	// An empty projected contribution is an explicit omission/replacement state;
+	// only a validated projection may make that context-edit-aware decision.
+	if (projection?.entries.some((entry) => entry.sourceEntry.id === entryId && entry.messages.length === 0)) return undefined;
+
 	const entry = entries.find((candidate) => candidate.id === entryId);
 	if (!entry) return undefined;
 	return findMessageIndexByRoleAndTimestamp(entry, messages);
@@ -2157,8 +2258,22 @@ export function rebuildDecisionStateFromEntries(entries: MinimalSessionEntry[]):
 // /prune picker helpers (pe-8re9)
 // ============================================================================
 
-/** Convert branch entries into the plain message array the pipeline/pickers correlate against. */
-export function sessionEntriesToMessages(entries: MinimalSessionEntry[]): MinimalMessage[] {
+function resolveSessionProjection(ctx: ExtensionContext): MinimalSessionProjection | undefined {
+	try {
+		const sessionManager = ctx.sessionManager as unknown as { buildSessionProjection?: () => unknown };
+		const buildProjection = sessionManager.buildSessionProjection;
+		if (typeof buildProjection !== "function") return undefined;
+		return normalizeSessionProjection(buildProjection.call(sessionManager));
+	} catch {
+		return undefined;
+	}
+}
+
+export function sessionEntriesToMessages(
+	entries: MinimalSessionEntry[],
+	projection?: MinimalSessionProjection,
+): MinimalMessage[] {
+	if (projection) return sessionProjectionToMessages(projection);
 	const messages: MinimalMessage[] = [];
 	for (const entry of entries) {
 		if (entry.type === "message" && entry.message) messages.push(entry.message);
@@ -2273,7 +2388,7 @@ export function formatPrunableItemDetail(item: PrunableItem, cost?: CacheCostMod
 
 /** Plain-text report used both by the non-UI /prune fallback and as a debugging aid. */
 export function formatPrunableItemsReport(items: PrunableItem[]): string[] {
-	if (items.length === 0) return ["No prunable tool results found in this session."];
+	if (items.length === 0) return ["No prunable tool results found in the current model context."];
 	return items.map((item, index) => `${index + 1}. ${formatPrunableItemOption(item)}`);
 }
 
@@ -2410,7 +2525,7 @@ export function formatStatusReport(input: StatusReportInput): string[] {
 	const lines: string[] = [];
 	lines.push(`Dynamic context pruning: ${config.enabled ? "ENABLED" : "disabled"}`);
 	lines.push(
-		`Gate: mode=${config.gate.mode} threshold=${config.gate.breakEvenThreshold} cachedPriceRatio=${config.gate.cachedPriceRatio}`,
+		`Gate: mode=${config.gate.mode} thresholds={idle:${resolveBreakEvenThreshold(config.gate, "idle")}, mid_loop:${resolveBreakEvenThreshold(config.gate, "mid_loop")}} cachedPriceRatio=${config.gate.cachedPriceRatio}`,
 	);
 	lines.push("Strategies:");
 	lines.push(`  dedupe: ${config.strategies.dedupe.enabled ? "on" : "off"}`);
@@ -2581,7 +2696,8 @@ export default function dynamicContextPruningExtension(pi: ExtensionAPI) {
 		description: "Review, manually prune, and restore prunable tool results (interactive picker in TUI/RPC)",
 		handler: async (_args, ctx) => {
 			const branchEntries = ctx.sessionManager.getBranch() as unknown as MinimalSessionEntry[];
-			const messages = sessionEntriesToMessages(branchEntries);
+			const projection = resolveSessionProjection(ctx);
+			const messages = sessionEntriesToMessages(branchEntries, projection);
 			const activeByToolCallId = buildActiveResultDecisionMap(persistedDecisions);
 			const restoredToolCallIds = new Set(
 				Array.from(lastDecisionByKey.values())
@@ -2591,19 +2707,20 @@ export default function dynamicContextPruningExtension(pi: ExtensionAPI) {
 			const items = buildPrunableItems(messages, activeByToolCallId, restoredToolCallIds);
 
 			if (!ctx.hasUI) {
-				notify(ctx, ["/prune: no interactive UI available. Prunable tool results in this session:", "", ...formatPrunableItemsReport(items)].join("\n"), "info");
+				notify(ctx, ["/prune: no interactive UI available. Prunable tool results in the current model context:", "", ...formatPrunableItemsReport(items)].join("\n"), "info");
 				return;
 			}
 
 			if (items.length === 0) {
-				notify(ctx, "No prunable tool results found in this session.", "info");
+				notify(ctx, "No prunable tool results found in the current model context.", "info");
 				return;
 			}
 
 			const doneLabel = "Done";
 			while (true) {
 				const latestBranch = ctx.sessionManager.getBranch() as unknown as MinimalSessionEntry[];
-				const latestMessages = sessionEntriesToMessages(latestBranch);
+				const latestProjection = resolveSessionProjection(ctx);
+				const latestMessages = sessionEntriesToMessages(latestBranch, latestProjection);
 				const latestActive = buildActiveResultDecisionMap(persistedDecisions);
 				const latestRestoredToolCallIds = new Set(
 					Array.from(lastDecisionByKey.values())
@@ -2612,7 +2729,7 @@ export default function dynamicContextPruningExtension(pi: ExtensionAPI) {
 				);
 				const latestItems = buildPrunableItems(latestMessages, latestActive, latestRestoredToolCallIds);
 				if (latestItems.length === 0) {
-					notify(ctx, "No prunable tool results found in this session.", "info");
+					notify(ctx, "No prunable tool results found in the current model context.", "info");
 					return;
 				}
 

package/extensions/dynamic-context-pruning/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-dynamic-context-pruning",
-  "version": "0.1.8",
+  "version": "0.1.10",
   "type": "module",
   "description": "A pi extension that dynamically prunes stale context to keep sessions lean.",
   "keywords": [

package/extensions/oracle/index.ts

@@ -107,6 +107,13 @@ const ORACLE_CONFIG_FILE = "oracle.json";
 
 const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 	"amazon-bedrock": [
+		"global.anthropic.claude-opus-5-5 ",
+		"au.anthropic.claude-opus-5-5 ",
+		"eu.anthropic.claude-opus-5-5 ",
+		"jp.anthropic.claude-opus-5-5 ",
+		"us.anthropic.claude-opus-5-5 ",
+		"anthropic.claude-opus-5-5 ",
+		"claude-opus-5-5",
 		"claude-opus-5",
 		"claude-fable-5",
 		"claude-opus-4-8",
@@ -115,6 +122,10 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"claude-opus-4-5",
 		"claude-opus-4-1",
 		"claude-opus-4",
+		"global.openai.gpt-6-astra ",
+		"us.openai.gpt-6-astra ",
+		"openai.gpt-6-astra ",
+		"gpt-6-astra",
 		"openai.gpt-5.6-sol",
 		"openai.gpt-5.6-terra",
 		"openai.gpt-5.6-luna",
@@ -123,7 +134,6 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"claude-sonnet-4-5",
 		"claude-sonnet-4",
 		"deepseek.v3.2",
-		"deepseek.r1",
 		"kimi-k2.5",
 		"minimax-m2.5",
 		"minimax-m2.1",
@@ -131,6 +141,8 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"xai.grok-4.6",
 	],
 	anthropic: [
+		"claude-opus-5-5 ",
+		"claude-opus-5-5",
 		"claude-opus-5",
 		"claude-fable-5",
 		"claude-opus-4-8",
@@ -143,8 +155,14 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 	],
 	"ant-ling": ["Ling-2.6-1T", "Ling-2.6-flash"],
 	"azure-openai-responses": [
+		"gpt-6-astra ",
+		"gpt-6-astra",
+		"gpt-6-sol ",
+		"gpt-6-sol",
 		"gpt-5.6-sol",
 		"gpt-5.6-terra",
+		"gpt-6-luna ",
+		"gpt-6-luna",
 		"gpt-5.6-luna",
 		"gpt-5.5-pro",
 		"gpt-5.5",
@@ -181,7 +199,7 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"zai-org/GLM-5.2-Fast",
 		"zai-org/GLM-4.7",
 	],
-	cerebras: ["gpt-oss-120b", "gemma-4-31b"],
+	cerebras: ["gpt-oss-120b", "qwen-3.8-27b"],
 	"cloudflare-ai-gateway": [
 		"claude-fable-5",
 		"claude-opus-5",
@@ -189,6 +207,8 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"claude-opus-4.7",
 		"claude-opus-4.6",
 		"claude-opus-4.5",
+		"gpt-6-astra ",
+		"gpt-6-astra",
 		"gpt-5.5",
 		"gpt-5.4",
 		"workers-ai/@cf/moonshotai/kimi-k2.7-code",
@@ -212,7 +232,7 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"@cf/qwen/qwen3-30b-a3b-fp8",
 		"@cf/zai-org/glm-4.7-flash",
 	],
-	deepseek: ["deepseek-v4-pro", "deepseek-v4-flash"],
+	deepseek: ["deepseek-v4-pro", "deepseek-flash"],
 	fireworks: [
 		"accounts/fireworks/models/kimi-k3",
 		"accounts/fireworks/routers/kimi-k3-fast",
@@ -226,13 +246,20 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"accounts/fireworks/models/gpt-oss-120b",
 	],
 	"github-copilot": [
+		"claude-opus-5.5 ",
+		"claude-opus-5.5",
 		"claude-opus-5",
 		"claude-fable-5",
 		"claude-opus-4.8",
 		"claude-opus-4.7",
+		"gpt-6-astra ",
 		"gpt-6-astra",
+		"gpt-6-sol ",
+		"gpt-6-sol",
 		"gpt-5.6-sol",
 		"gpt-5.6-terra",
+		"gpt-6-luna ",
+		"gpt-6-luna",
 		"gpt-5.6-luna",
 		"gpt-5.5",
 		"gpt-5.4",
@@ -241,6 +268,10 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"claude-sonnet-5",
 		"claude-sonnet-4.6",
 		"gpt-5-mini",
+		"grok-4.7 ",
+		"grok-4.7",
+		"grok-4.6",
+		"grok-4.5",
 		"kimi-k3",
 		"kimi-k2.7-code",
 	],
@@ -285,6 +316,13 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"Qwen/Qwen3-Coder-Next",
 	],
 	"kimi-coding": ["k3", "kimi-for-coding", "kimi-for-coding-highspeed"],
+	meta: [
+		"muse-spark-1.3",
+		"muse-spark-1.3-contributor",
+		"muse-spark-1.2",
+		"muse-spark-1.2-contributor",
+		"muse-spark-1.1",
+	],
 	minimax: ["MiniMax-M3", "MiniMax-M2.7-highspeed", "MiniMax-M2.7"],
 	"minimax-cn": ["MiniMax-M3", "MiniMax-M2.7-highspeed", "MiniMax-M2.7"],
 	mistral: [
@@ -301,9 +339,14 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"zai-glm-5-2",
 	],
 	openai: [
+		"gpt-6-astra ",
 		"gpt-6-astra",
+		"gpt-6-sol ",
+		"gpt-6-sol",
 		"gpt-5.6-sol",
 		"gpt-5.6-terra",
+		"gpt-6-luna ",
+		"gpt-6-luna",
 		"gpt-5.6-luna",
 		"gpt-5.5-pro",
 		"gpt-5.5",
@@ -321,18 +364,32 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"gpt-5-mini",
 	],
 	"openai-codex": [
+		"gpt-6-astra ",
 		"gpt-6-astra",
+		"gpt-6-sol ",
+		"gpt-6-sol",
 		"gpt-5.6-sol",
 		"gpt-5.6-terra",
+		"gpt-6-luna ",
+		"gpt-6-luna",
 		"gpt-5.6-luna",
 		"gpt-5.5",
-		"gpt-5.4",
-		"gpt-5.4-mini",
 		"gpt-5.3-codex-spark",
 	],
 	opencode: [
+		"claude-opus-5-5 ",
+		"claude-opus-5-5",
 		"claude-opus-5",
 		"claude-fable-5",
+		"gpt-6-astra ",
+		"gpt-6-astra",
+		"gpt-6-sol ",
+		"gpt-6-sol",
+		"gpt-5.6-sol",
+		"gpt-5.6-terra",
+		"gpt-6-luna ",
+		"gpt-6-luna",
+		"gpt-5.6-luna",
 		"gpt-5.5-pro",
 		"gpt-5.5",
 		"gpt-5.4-pro",
@@ -344,6 +401,7 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"gpt-5.3-codex",
 		"gpt-5.2-codex",
 		"gpt-5.1-codex",
+		"grok-4.6",
 		"grok-4.5",
 		"gemini-3.5-flash",
 		"gemini-3.1-pro",
@@ -374,8 +432,13 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"kimi-k3",
 		"kimi-k2.7-code",
 		"kimi-k2.6",
+		"grok-4.7 ",
+		"grok-4.7",
+		"grok-4.6",
 	],
 	openrouter: [
+		"anthropic/claude-opus-5.5 ",
+		"anthropic/claude-opus-5.5",
 		"anthropic/claude-opus-5",
 		"anthropic/claude-fable-5",
 		"~anthropic/claude-fable-latest",
@@ -384,15 +447,21 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"anthropic/claude-opus-4.6",
 		"anthropic/claude-opus-4.5",
 		"anthropic/claude-opus-4.1",
+		"openai/gpt-6-astra ",
+		"openai/gpt-6-astra",
+		"openai/gpt-6-sol ",
+		"openai/gpt-6-sol",
 		"openai/gpt-5.6-sol-pro",
 		"openai/gpt-5.6-sol",
 		"openai/gpt-5.6-terra-pro",
 		"openai/gpt-5.6-terra",
+		"openai/gpt-6-luna ",
+		"openai/gpt-6-luna",
 		"openai/gpt-5.6-luna-pro",
 		"openai/gpt-5.6-luna",
 		"openai/gpt-5.5-pro",
 		"openai/gpt-5.5",
-		"~openai/gpt-latest",
+		"~openai/gpt-astra-latest",
 		"google/gemini-3.1-pro-preview-customtools",
 		"google/gemini-3.1-pro-preview",
 		"~google/gemini-pro-latest",
@@ -403,6 +472,9 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"~moonshotai/kimi-latest",
 		"moonshotai/kimi-k2.6",
 		"moonshotai/kimi-k2-thinking",
+		"x-ai/grok-4.7 ",
+		"x-ai/grok-4.7",
+		"x-ai/grok-4.6",
 		"x-ai/grok-4.5",
 		"~x-ai/grok-latest",
 		"deepseek/deepseek-v4-pro",
@@ -444,6 +516,34 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"glm-5.2",
 		"deepseek-v4-flash-0731",
 	],
+	radius: [
+		"gpt-6-astra ",
+		"gpt-6-astra",
+		"claude-opus-5-5 ",
+		"claude-opus-5-5",
+		"claude-opus-5",
+		"gpt-6-sol ",
+		"gpt-6-sol",
+		"gpt-5.6-sol",
+		"gpt-5.6-terra",
+		"gpt-6-luna ",
+		"gpt-6-luna",
+		"gpt-5.6-luna",
+		"claude-fable-5-1",
+		"claude-fable-5",
+		"claude-opus-4-8",
+		"claude-sonnet-5",
+		"gpt-5.5",
+		"gpt-5.4",
+		"gpt-5.3-codex",
+		"kimi-k3",
+		"deepseek-v4-pro",
+		"glm-5.3",
+		"kimi-k2.7-code",
+		"deepseek-v4.1-flash",
+		"gpt-5.4-mini",
+		"glm-5.3-flash",
+	],
 	together: [
 		"deepseek-ai/DeepSeek-V4-Pro",
 		"moonshotai/Kimi-K3",
@@ -461,6 +561,8 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"google/gemma-4-31B-it",
 	],
 	"vercel-ai-gateway": [
+		"anthropic/claude-opus-5.5 ",
+		"anthropic/claude-opus-5.5",
 		"anthropic/claude-opus-5",
 		"anthropic/claude-fable-5",
 		"anthropic/claude-opus-4.8",
@@ -469,8 +571,14 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"anthropic/claude-opus-4.5",
 		"anthropic/claude-sonnet-5",
 		"anthropic/claude-sonnet-4.6",
+		"openai/gpt-6-astra ",
+		"openai/gpt-6-astra",
+		"openai/gpt-6-sol ",
+		"openai/gpt-6-sol",
 		"openai/gpt-5.6-sol",
 		"openai/gpt-5.6-terra",
+		"openai/gpt-6-luna ",
+		"openai/gpt-6-luna",
 		"openai/gpt-5.6-luna",
 		"openai/gpt-5.5-pro",
 		"openai/gpt-5.5",
@@ -495,6 +603,9 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"google/gemini-3.1-pro-preview",
 		"google/gemini-3.5-flash",
 		"google/gemini-3-flash",
+		"spacexai/grok-4.7 ",
+		"spacexai/grok-4.7",
+		"spacexai/grok-4.6",
 		"xai/grok-4.5",
 		"xai/grok-4.20-reasoning",
 		"minimax/minimax-m3",
@@ -503,6 +614,8 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"zai/glm-5.1",
 	],
 	xai: [
+		"grok-4.7 ",
+		"grok-4.7",
 		"grok-4.6",
 		"grok-4.5",
 		"grok-4.3",
@@ -515,9 +628,7 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 		"nvidia/nemotron-3-ultra-550b-a55b",
 		"nvidia/nemotron-3-super-120b-a12b",
 		"moonshotai/kimi-k3",
-		"deepseek-ai/deepseek-v4-pro-0813",
 		"moonshotai/kimi-k2.6",
-		"minimaxai/minimax-m3",
 		"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning",
 	],
 	zai: [
@@ -528,13 +639,12 @@ const PROVIDER_MODEL_PREFERENCES: Record<string, string[]> = {
 	],
 	"zai-coding-cn": [
 		"glm-5.3",
-		"glm-5.2",
-		"glm-5-turbo",
-		"glm-5",
-		"glm-4.7",
+		"glm-5.3-highspeed",
+		"glm-5.3-flash",
+		"glm-4.6v",
 	],
-	moonshotai: ["kimi-k3", "kimi-k2.7-code", "kimi-k2.7-code-highspeed", "kimi-k2.6", "kimi-k2-thinking-turbo", "kimi-k2-thinking", "kimi-k2.5"],
-	"moonshotai-cn": ["kimi-k3", "kimi-k2.7-code", "kimi-k2.7-code-highspeed", "kimi-k2.6", "kimi-k2-thinking-turbo", "kimi-k2-thinking", "kimi-k2.5"],
+	moonshotai: ["kimi-k3", "kimi-k2.7-code", "kimi-k2.7-code-highspeed", "kimi-k2.6"],
+	"moonshotai-cn": ["kimi-k3", "kimi-k2.7-code", "kimi-k2.7-code-highspeed", "kimi-k2.6"],
 };
 
 const ORACLE_SYSTEM_PROMPT = [
@@ -913,7 +1023,7 @@ function clampThinkingLevel(model: PiModel, requested: ThinkingLevel): ThinkingL
 
 function isGpt56SolModel(model: PiModel | undefined): boolean {
 	if (!model) return false;
-	return /\bgpt-5\.6-sol\b/i.test(`${model.id} ${model.name ?? ""}`);
+	return /\bgpt-(?:5\.6|6)-sol\b/i.test(`${model.id} ${model.name ?? ""}`);
 }
 
 function resolveThinkingLevel(

package/extensions/oracle/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@diegopetrucci/pi-oracle",
-  "version": "0.1.27",
+  "version": "0.1.28",
   "description": "An Amp-style oracle extension for pi that consults the strongest reasoning model on your current provider.",
   "keywords": [
     "pi-package",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@diegopetrucci/pi-extensions",
-  "version": "0.1.67",
+  "version": "0.1.69",
   "description": "A collection of pi extensions and skills for annotation UIs, context management, workflow audits, contrarian review, review-comment triage, notifications, brrr push alerts, safety guards, GitHub research, repo-local knowledge, todos, tool rendering, model/provider helpers, and Xiaohei-style article illustrations.",
   "keywords": [
     "pi-package",
@@ -71,9 +71,9 @@
     ]
   },
   "devDependencies": {
-    "@earendil-works/pi-ai": "^0.85.1",
-    "@earendil-works/pi-coding-agent": "^0.85.1",
-    "@earendil-works/pi-tui": "^0.85.1",
+    "@earendil-works/pi-ai": "^0.87.1",
+    "@earendil-works/pi-coding-agent": "^0.87.1",
+    "@earendil-works/pi-tui": "^0.87.1",
     "@types/node": "^25.9.1",
     "husky": "^9.1.7",
     "typebox": "^1.3.6",