feat(models): add GPT-6.1 Sol catalogs, pricing, and Copilot routing

Dockerfile

@@ -363,7 +363,7 @@ RUN --mount=type=cache,id=s/92ca8a61-c1ba-421f-a389-d48ac7258c2d-apt-cache,targe
 #      build, not the floating `@latest`.
 RUN --mount=type=cache,id=s/92ca8a61-c1ba-421f-a389-d48ac7258c2d-npm-cache,target=/root/.npm \
   npm install -g --no-audit --no-fund \
-    @openai/codex@0.156.1 \
+    @openai/codex@0.159.2 \
     @anthropic-ai/claude-code@2.1.260 \
     droid@0.212.0 \
     openclaw@2026.9.1

open-sse/config/providers/registry/codex/index.ts

@@ -66,6 +66,15 @@ export const codexProvider: RegistryEntry = {
     },
     { id: "gpt-6-sol-medium", name: "GPT 6 Sol (Medium)", ...GPT_5_6_CODEX_CAPABILITIES },
     { id: "gpt-6-sol-low", name: "GPT 6 Sol (Low)", ...GPT_5_6_CODEX_CAPABILITIES },
+    // GPT-6.1 Sol: low..ultra effort aliases, same windows as Sol.
+    // https://developers.openai.com/api/docs/models/gpt-6.1-sol
+    { id: "gpt-6.1-sol", name: "GPT 6.1 Sol", ...GPT_5_6_CODEX_CAPABILITIES },
+    { id: "gpt-6.1-sol-ultra", name: "GPT 6.1 Sol (Ultra)", ...GPT_5_6_CODEX_CAPABILITIES },
+    { id: "gpt-6.1-sol-max", name: "GPT 6.1 Sol (Max)", ...GPT_5_6_CODEX_CAPABILITIES },
+    { id: "gpt-6.1-sol-xhigh", name: "GPT 6.1 Sol (xHigh)", ...GPT_5_6_CODEX_CAPABILITIES, timeoutMs: 1200000 },
+    { id: "gpt-6.1-sol-high", name: "GPT 6.1 Sol (High)", ...GPT_5_6_CODEX_CAPABILITIES, timeoutMs: 1200000 },
+    { id: "gpt-6.1-sol-medium", name: "GPT 6.1 Sol (Medium)", ...GPT_5_6_CODEX_CAPABILITIES },
+    { id: "gpt-6.1-sol-low", name: "GPT 6.1 Sol (Low)", ...GPT_5_6_CODEX_CAPABILITIES },
     { id: "gpt-6-luna", name: "GPT 6 Luna", ...GPT_5_6_CODEX_CAPABILITIES },
     { id: "gpt-6-luna-max", name: "GPT 6 Luna (Max)", ...GPT_5_6_CODEX_CAPABILITIES },
     {

open-sse/config/providers/registry/ghe-copilot/index.ts

@@ -124,6 +124,24 @@ export const gheCopilotProvider: RegistryEntry = {
       contextLength: 1000000,
       maxOutputTokens: 64000,
     },
+    {
+      id: "gpt-6-astra",
+      name: "GPT-6 Astra",
+      supportsReasoning: true,
+      supportedThinkingEfforts: ["low", "medium", "high", "xhigh", "max", "ultra"],
+      targetFormat: "openai-responses",
+      contextLength: 1050000,
+      maxOutputTokens: 128000,
+    },
+    {
+      id: "gpt-6.1-sol",
+      name: "GPT-6.1 Sol",
+      supportsReasoning: true,
+      supportedThinkingEfforts: ["low", "medium", "high", "xhigh", "max", "ultra"],
+      targetFormat: "openai-responses",
+      contextLength: 1050000,
+      maxOutputTokens: 128000,
+    },
     {
       id: "gpt-5.6-sol",
       name: "GPT-5.6 Sol",

open-sse/config/providers/registry/github/index.ts

@@ -160,6 +160,24 @@ export const githubProvider: RegistryEntry = {
       contextLength: 1000000,
       maxOutputTokens: 64000,
     },
+    {
+      id: "gpt-6-astra",
+      name: "GPT-6 Astra",
+      supportsReasoning: true,
+      supportedThinkingEfforts: ["low", "medium", "high", "xhigh", "max", "ultra"],
+      targetFormat: "openai-responses",
+      contextLength: 1050000,
+      maxOutputTokens: 128000,
+    },
+    {
+      id: "gpt-6.1-sol",
+      name: "GPT-6.1 Sol",
+      supportsReasoning: true,
+      supportedThinkingEfforts: ["low", "medium", "high", "xhigh", "max", "ultra"],
+      targetFormat: "openai-responses",
+      contextLength: 1050000,
+      maxOutputTokens: 128000,
+    },
     {
       id: "gpt-5.6-sol",
       name: "GPT-5.6 Sol",

open-sse/executors/codex/reasoningSuffix.ts

@@ -15,12 +15,14 @@ export const CODEX_MAX_ALIAS_MODELS = new Set([
   "gpt-6-astra",
   "gpt-6-sol",
   "gpt-6-luna",
+  "gpt-6.1-sol",
 ]);
 export const CODEX_ULTRA_ALIAS_MODELS = new Set([
   "gpt-5.6-sol",
   "gpt-5.6-terra",
   "gpt-6-astra",
   "gpt-6-sol",
+  "gpt-6.1-sol",
 ]);
 
 /** Highest effort a max/ultra-tier base model accepts, or null for other models. */

open-sse/executors/github.ts

@@ -110,7 +110,7 @@ export class GithubExecutor extends BaseExecutor {
     // 9router#1536: but never route Gemini/Claude variants to /responses (they 400) —
     // gate the whole decision on supportsResponsesEndpoint().
     if (
-      (targetFormat === "openai-responses" || /codex/i.test(model)) &&
+      (targetFormat === "openai-responses" || /codex|^gpt-6/i.test(model)) &&
       this.supportsResponsesEndpoint(model)
     ) {
       return (

open-sse/translator/request/openai-responses/helpers.ts

@@ -51,7 +51,7 @@ export function imageUrlToText(value: unknown): string {
 }
 
 const CODEX_MAX_EFFORT_MODEL_PATTERN =
-  /^(?:gpt-5\.6-(?:sol|terra|luna)|gpt-6-(?:astra|sol|luna))(?:-(?:none|low|medium|high|xhigh|max|ultra))?$/;
+  /^(?:gpt-5\.6-(?:sol|terra|luna)|gpt-6(?:\.\d+)?-(?:astra|sol|luna))(?:-(?:none|low|medium|high|xhigh|max|ultra))?$/;
 const KIRO_GPT_5_6_MODEL_PATTERN =
   /^(?:kiro|kr)\/gpt-5\.6-(?:sol|terra|luna)(?:-(?:none|low|medium|high|xhigh|max))?$/;
 

src/lib/providers/codexFastTier.ts

@@ -16,6 +16,7 @@ export type CodexGlobalServiceMode = "none" | CodexServiceTier;
 export const CODEX_FAST_TIER_DEFAULT_SUPPORTED_MODELS: readonly string[] = [
   "gpt-6-astra",
   "gpt-6-sol",
+  "gpt-6.1-sol",
   "gpt-6-luna",
   "gpt-5.6-sol",
   "gpt-5.6-terra",

src/lib/usage/costCalculator.ts

@@ -103,8 +103,8 @@ export function getCodexFastCostMultiplier(
   const compactModelKey = modelKey.replace(/-/g, "");
   // Codex GPT-6 Fast is 2.5x Standard (https://developers.openai.com/codex/pricing).
   if (
-    /^gpt-6-(?:astra|sol|luna)$/.test(modelKey) ||
-    /^gpt6(?:astra|sol|luna)$/.test(compactModelKey)
+    /^gpt-6(?:\.\d+)?-(?:astra|sol|luna)$/.test(modelKey) ||
+    /^gpt6(?:\d+)?(?:astra|sol|luna)$/.test(compactModelKey)
   ) {
     return 2.5;
   }
@@ -119,6 +119,19 @@ export function getCodexFastCostMultiplier(
   return 1;
 }
 
+const GPT61_LONG_CONTEXT_INPUT_TOKENS = 272_000;
+
+function longContextPriceMultiplier(
+  model: string | null | undefined,
+  inputTokens: number
+): { input: number; cache: number; output: number } {
+  const flat = { input: 1, cache: 1, output: 1 };
+  if (inputTokens <= GPT61_LONG_CONTEXT_INPUT_TOKENS) return flat;
+  const modelKey = stripCodexEffortSuffix(normalizeModelName(String(model || "")).toLowerCase());
+  if (!/^gpt-6\.\d+-sol$/.test(modelKey)) return flat;
+  return { input: 2, cache: 2, output: 1.5 };
+}
+
 /**
  * Calculate cost for a usage entry.
  *
@@ -162,20 +175,26 @@ export function computeCostFromPricing(
   // so we must subtract BOTH cache types to avoid pricing cache at the full
   // input rate in addition to their dedicated cache_* rates below.
   const nonCachedInput = Math.max(0, inputTokens - cachedTokens - cacheCreationTokens);
-  cost += nonCachedInput * (inputPrice / 1_000_000);
-  if (cachedTokens > 0) cost += cachedTokens * (cachedPrice / 1_000_000);
+  // GPT-6.1 bills the whole request at the long-context rate once input
+  // crosses 272K: 2x input and cache, 1.5x output.
+  // https://developers.openai.com/api/docs/models/gpt-6.1-sol
+  const longContext = longContextPriceMultiplier(options.model, inputTokens);
+  cost += nonCachedInput * ((inputPrice * longContext.input) / 1_000_000);
+  if (cachedTokens > 0) cost += cachedTokens * ((cachedPrice * longContext.cache) / 1_000_000);
 
   const outputTokens = tokens.output ?? tokens.completion_tokens ?? tokens.output_tokens ?? 0;
-  cost += outputTokens * (outputPrice / 1_000_000);
+  cost += outputTokens * ((outputPrice * longContext.output) / 1_000_000);
 
   // completion_tokens is reasoning-inclusive. Reasoning is already billed at
   // the output rate above, so a dedicated price contributes only its premium.
   const reasoningTokens = tokens.reasoning ?? tokens.reasoning_tokens ?? 0;
   if (reasoningTokens > 0 && pricing.reasoning !== undefined && pricing.reasoning !== null) {
-    cost += reasoningTokens * ((reasoningPrice - outputPrice) / 1_000_000);
+    cost += reasoningTokens * (Math.max(0, reasoningPrice - outputPrice * longContext.output) / 1_000_000);
   }
 
-  if (cacheCreationTokens > 0) cost += cacheCreationTokens * (cacheCreationPrice / 1_000_000);
+  if (cacheCreationTokens > 0) {
+    cost += cacheCreationTokens * ((cacheCreationPrice * longContext.cache) / 1_000_000);
+  }
 
   return cost * getCodexFastCostMultiplier(options.provider, options.model, options.serviceTier);
 }

src/shared/constants/codexClient.ts

@@ -3,7 +3,7 @@
 // refresh this so the fingerprint OpenAI sees from the OAuth/Responses face
 // matches the real client version. Overridable per-deployment via
 // CODEX_CLIENT_VERSION.
-export const DEFAULT_CODEX_CLIENT_VERSION = "0.156.1";
+export const DEFAULT_CODEX_CLIENT_VERSION = "0.159.2";
 export const CODEX_CLI_RS_ORIGINATOR = "codex_cli_rs";
 
 export function getCodexCliRsHeaders(

src/shared/constants/pricing/oauth-subscriptions.ts

@@ -31,6 +31,16 @@ const GPT_6_SOL_CODEX_PRICING = {
   reasoning: 10.0,
   cache_creation: 2.5,
 };
+// GPT-6.1 Sol Standard, from models.dev openai/gpt-6.1-sol and the OpenAI
+// model page. Cached input is $0.10, not the $0.20 GPT-6 Sol rate.
+// https://developers.openai.com/api/docs/models/gpt-6.1-sol
+const GPT_6_1_SOL_CODEX_PRICING = {
+  input: 2.0,
+  output: 10.0,
+  cached: 0.1,
+  reasoning: 10.0,
+  cache_creation: 2.5,
+};
 const GPT_6_LUNA_CODEX_PRICING = {
   input: 0.1,
   output: 0.5,
@@ -122,6 +132,13 @@ export const DEFAULT_PRICING_OAUTH = {
     "gpt-6-sol-high": GPT_6_SOL_CODEX_PRICING,
     "gpt-6-sol-medium": GPT_6_SOL_CODEX_PRICING,
     "gpt-6-sol-low": GPT_6_SOL_CODEX_PRICING,
+    "gpt-6.1-sol": GPT_6_1_SOL_CODEX_PRICING,
+    "gpt-6.1-sol-ultra": GPT_6_1_SOL_CODEX_PRICING,
+    "gpt-6.1-sol-max": GPT_6_1_SOL_CODEX_PRICING,
+    "gpt-6.1-sol-xhigh": GPT_6_1_SOL_CODEX_PRICING,
+    "gpt-6.1-sol-high": GPT_6_1_SOL_CODEX_PRICING,
+    "gpt-6.1-sol-medium": GPT_6_1_SOL_CODEX_PRICING,
+    "gpt-6.1-sol-low": GPT_6_1_SOL_CODEX_PRICING,
     "gpt-6-luna": GPT_6_LUNA_CODEX_PRICING,
     "gpt-6-luna-max": GPT_6_LUNA_CODEX_PRICING,
     "gpt-6-luna-xhigh": GPT_6_LUNA_CODEX_PRICING,

src/shared/reasoning/effortStandardization.ts

@@ -41,7 +41,7 @@ export function extendCodexGpt56EffortValues(
   }
 
   const match = normalizedModel.match(
-    /^gpt-(?:5\.6-(sol|terra|luna)|6-(astra|sol|luna))(?:-(?:none|low|medium|high|xhigh|max|ultra))?$/
+    /^gpt-(?:5\.6-(sol|terra|luna)|6(?:\.\d+)?-(astra|sol|luna))(?:-(?:none|low|medium|high|xhigh|max|ultra))?$/
   );
   if (!match) return values;
 
@@ -53,7 +53,8 @@ export function extendCodexGpt56EffortValues(
   if (normalizedProvider !== "codex" && normalizedProvider !== "cx") return values;
 
   const nativeValues = ["low", "medium", "high", "xhigh", "max"];
-  return (match[1] || match[2]) === "luna" ? nativeValues : [...nativeValues, "ultra"];
+  const family = match[1] || match[2];
+  return family === "luna" ? nativeValues : [...nativeValues, "ultra"];
 }
 
 /**

tests/snapshots/provider/translate-path.json

@@ -1334,23 +1334,23 @@
         "Authorization": "Bearer <TOK>",
         "Content-Type": "application/json",
         "Openai-Beta": "responses_websockets=2026-02-06",
-        "User-Agent": "codex-cli/0.156.1 (<OS>; <ARCH>)",
-        "Version": "0.156.1"
+        "User-Agent": "codex-cli/0.159.2 (<OS>; <ARCH>)",
+        "Version": "0.159.2"
       },
       "nonStream": {
         "Authorization": "Bearer <TOK>",
         "Content-Type": "application/json",
         "Openai-Beta": "responses_websockets=2026-02-06",
-        "User-Agent": "codex-cli/0.156.1 (<OS>; <ARCH>)",
-        "Version": "0.156.1"
+        "User-Agent": "codex-cli/0.159.2 (<OS>; <ARCH>)",
+        "Version": "0.159.2"
       },
       "oauth": {
         "Accept": "text/event-stream",
         "Authorization": "Bearer <TOK>",
         "Content-Type": "application/json",
         "Openai-Beta": "responses_websockets=2026-02-06",
-        "User-Agent": "codex-cli/0.156.1 (<OS>; <ARCH>)",
-        "Version": "0.156.1"
+        "User-Agent": "codex-cli/0.159.2 (<OS>; <ARCH>)",
+        "Version": "0.159.2"
       }
     },
     "url": {

tests/unit/8951-github-gpt56-responses.test.ts

@@ -6,14 +6,15 @@ import { GithubExecutor } from "../../open-sse/executors/github.ts";
 test("#8951 GitHub GPT-5.6 models must use the Responses endpoint", () => {
   const executor = new GithubExecutor();
   const urls = Object.fromEntries(
-    ["gpt-5.6-sol", "gpt-5.6-terra", "gpt-5.6-luna"].map((model) => [
-      model,
-      executor.buildUrl(model, false),
-    ])
+    ["gpt-5.6-sol", "gpt-5.6-terra", "gpt-5.6-luna", "gpt-6-astra", "gpt-6.1-sol"].map(
+      (model) => [model, executor.buildUrl(model, false)]
+    )
   );
   assert.deepEqual(urls, {
     "gpt-5.6-sol": "https://api.githubcopilot.com/responses",
     "gpt-5.6-terra": "https://api.githubcopilot.com/responses",
     "gpt-5.6-luna": "https://api.githubcopilot.com/responses",
+    "gpt-6-astra": "https://api.githubcopilot.com/responses",
+    "gpt-6.1-sol": "https://api.githubcopilot.com/responses",
   });
 });

tests/unit/codex-fast-tier.test.ts

@@ -48,6 +48,7 @@ test("Codex global service mode distinguishes no setting from explicit tiers", (
       supportedModels: [
         "gpt-6-astra",
         "gpt-6-sol",
+        "gpt-6.1-sol",
         "gpt-6-luna",
         "gpt-5.6-sol",
         "gpt-5.6-terra",

tests/unit/codex-gpt6-sol-luna.test.ts

@@ -5,14 +5,15 @@ import { getModelsByProviderId } from "../../open-sse/config/providerModels.ts";
 import { CodexExecutor } from "../../open-sse/executors/codex.ts";
 import { openaiToOpenAIResponsesRequest } from "../../open-sse/translator/request/openai-responses/toResponses.ts";
 import { getPricingForModel } from "../../src/shared/constants/pricing.ts";
-import { getCodexFastCostMultiplier } from "../../src/lib/usage/costCalculator.ts";
+import { computeCostFromPricing, getCodexFastCostMultiplier } from "../../src/lib/usage/costCalculator.ts";
 import { extendCodexGpt56EffortValues } from "../../src/shared/reasoning/effortStandardization.ts";
 import * as reasoningMetadata from "../../src/lib/vscode/reasoningMetadata.ts";
 
 // The live Codex catalog (client 0.155.1) advertises Sol at low..ultra and Luna at
 // low..max, mirroring the GPT-5.6 Sol/Luna split.
 const FAMILIES = [
   { model: "gpt-6-sol", efforts: ["ultra", "max", "xhigh", "high", "medium", "low"] },
+  { model: "gpt-6.1-sol", efforts: ["ultra", "max", "xhigh", "high", "medium", "low"] },
   { model: "gpt-6-luna", efforts: ["max", "xhigh", "high", "medium", "low"] },
 ] as const;
 
@@ -112,6 +113,10 @@ test("Catalog effort tiers follow the live Codex levels for GPT-6 models", () =>
   for (const provider of ["codex", "cx"]) {
     assert.deepEqual(extendCodexGpt56EffortValues(provider, "gpt-6-astra", base), withUltra);
     assert.deepEqual(extendCodexGpt56EffortValues(provider, "gpt-6-sol", base), withUltra);
+    assert.deepEqual(extendCodexGpt56EffortValues(provider, "gpt-6.1-sol", base), withUltra);
+    // A later minor of the same family keeps ultra without a new allowlist entry.
+    assert.deepEqual(extendCodexGpt56EffortValues(provider, "gpt-6.2-sol", base), withUltra);
+    assert.deepEqual(extendCodexGpt56EffortValues(provider, "gpt-6.1-luna", base), withMax);
     assert.deepEqual(extendCodexGpt56EffortValues(provider, "gpt-6-luna-max", base), withMax);
   }
   // Kiro's GPT-5.6 max extension does not reach models Kiro does not serve.
@@ -179,6 +184,7 @@ test("GPT-6 Sol and Luna Codex pricing and Fast multiplier match the credit rate
   // Sol 50 / 5 / 250, Luna 2.5 / 0.25 / 12.5. Fast is 2.5x Standard for both.
   const expected = {
     "gpt-6-sol": { input: 2, cached: 0.2, output: 10 },
+    "gpt-6.1-sol": { input: 2, cached: 0.1, output: 10 },
     "gpt-6-luna": { input: 0.1, cached: 0.01, output: 0.5 },
   };
   for (const { model, efforts } of FAMILIES) {
@@ -196,3 +202,36 @@ test("GPT-6 Sol and Luna Codex pricing and Fast multiplier match the credit rate
     }
   }
 });
+
+test("GPT-6.1 Sol bills the whole request at the long-context rate past 272K input", () => {
+  const pricing = { input: 2, cached: 0.1, output: 10, cache_creation: 2.5, reasoning: 10 };
+  const tokens = { input: 300_000, cache_read_input_tokens: 50_000, output: 1_000 };
+  const standard = computeCostFromPricing(
+    pricing,
+    { input: 272_000, output: 1_000 },
+    { model: "gpt-6.1-sol" }
+  );
+  const elevated = computeCostFromPricing(
+    pricing,
+    { ...tokens, reasoning_tokens: 200 },
+    { model: "gpt-6.1-sol-high" }
+  );
+  const olderSol = computeCostFromPricing(pricing, tokens, { model: "gpt-6-sol" });
+
+  const standardExpected = (272_000 * 2 + 1_000 * 10) / 1_000_000;
+  // Output is already billed at 1.5x, which is above the reasoning price.
+  // The premium cannot go negative or the request would be under-charged.
+  const elevatedExpected = (250_000 * 4 + 50_000 * 0.2 + 1_000 * 15) / 1_000_000;
+  const olderExpected = (250_000 * 2 + 50_000 * 0.1 + 1_000 * 10) / 1_000_000;
+  assert.ok(Math.abs(standard - standardExpected) < 1e-9);
+  assert.ok(Math.abs(elevated - elevatedExpected) < 1e-9);
+  assert.ok(Math.abs(olderSol - olderExpected) < 1e-9);
+});
+
+test("GPT-6.1 Sol stays at the standard rate one token under 272K and doubles one token over", () => {
+  const pricing = { input: 2, output: 10 };
+  const under = computeCostFromPricing(pricing, { input: 271_999, output: 0 }, { model: "gpt-6.1-sol" });
+  const over = computeCostFromPricing(pricing, { input: 272_001, output: 0 }, { model: "gpt-6.1-sol" });
+  assert.ok(Math.abs(under - (271_999 * 2) / 1_000_000) < 1e-9);
+  assert.ok(Math.abs(over - (272_001 * 4) / 1_000_000) < 1e-9);
+});

tests/unit/executor-codex.test.ts

@@ -184,10 +184,10 @@ test("CodexExecutor.buildHeaders binds workspace ids and disables SSE accept for
   assert.equal(standardHeaders.Authorization, "Bearer codex-token");
   assert.equal(standardHeaders.Accept, "text/event-stream");
   assert.equal(standardHeaders["chatgpt-account-id"], "workspace-1");
-  assert.equal(standardHeaders.Version, "0.156.1");
+  assert.equal(standardHeaders.Version, "0.159.2");
   assert.equal(standardHeaders["Openai-Beta"], "responses_websockets=2026-02-06");
   assert.equal(standardHeaders["X-Codex-Beta-Features"], undefined);
-  assert.equal(standardHeaders["User-Agent"], "codex-cli/0.156.1 (Windows 10.0.26200; x64)");
+  assert.equal(standardHeaders["User-Agent"], "codex-cli/0.159.2 (Windows 10.0.26200; x64)");
   assert.equal(compactHeaders.Accept, "application/json");
 });
 
@@ -213,7 +213,7 @@ test("CodexExecutor.buildHeaders honors safe env overrides for Version and User-
     },
     () => {
       const headers = executor.buildHeaders({ accessToken: "codex-token" }, true);
-      assert.equal(headers.Version, "0.156.1");
+      assert.equal(headers.Version, "0.159.2");
       assert.equal(headers["User-Agent"], "custom-codex/9.9.9");
     }
   );