@usabledev/usable-chat 1.198.3 → 1.199.0

package/cli.js

@@ -45158,6 +45158,11 @@ var init_model_pricing = __esm({
         output: 5e-5,
         multiplier: 80
       },
+      "anthropic/claude-fable-5.1": {
+        input: 1e-5,
+        output: 5e-5,
+        multiplier: 80
+      },
       "anthropic/claude-haiku-4.5": {
         input: 1e-6,
         output: 5e-6,
@@ -45270,16 +45275,31 @@ var init_model_pricing = __esm({
         multiplier: 1
       },
       "deepseek/deepseek-v4-flash-0731": {
-        input: 8e-8,
+        input: 6e-8,
         output: 18e-8,
         multiplier: 1
       },
+      "deepseek/deepseek-v4-flash-0731:batch": {
+        input: 14e-8,
+        output: 28e-8,
+        multiplier: 1
+      },
+      "deepseek/deepseek-v4-flash-vision-exp": {
+        input: 2156e-10,
+        output: 6468e-10,
+        multiplier: 1
+      },
       "deepseek/deepseek-v4-pro": {
-        input: 13e-7,
-        output: 26e-7,
-        multiplier: 5
+        input: 12e-7,
+        output: 12e-7,
+        multiplier: 3
       },
       "deepseek/deepseek-v4-pro-0813": {
+        input: 5808e-10,
+        output: 17424e-10,
+        multiplier: 3
+      },
+      "deepseek/deepseek-v4-pro-0813:batch": {
         input: 132e-8,
         output: 396e-8,
         multiplier: 7
@@ -45431,9 +45451,19 @@ var init_model_pricing = __esm({
         multiplier: 6
       },
       "google/gemini-3.7-flash:batch": {
-        input: 1875e-10,
-        output: 9375e-10,
-        multiplier: 2
+        input: 375e-9,
+        output: 1875e-9,
+        multiplier: 3
+      },
+      "google/gemini-3.8-flash": {
+        input: 75e-8,
+        output: 375e-8,
+        multiplier: 6
+      },
+      "google/gemini-3.8-flash:batch": {
+        input: 375e-9,
+        output: 1875e-9,
+        multiplier: 3
       },
       "google/gemma-3-12b-it": {
         input: 5e-8,
@@ -45450,11 +45480,6 @@ var init_model_pricing = __esm({
         output: 1e-7,
         multiplier: 1
       },
-      "google/gemma-3n-e4b-it": {
-        input: 6e-8,
-        output: 12e-8,
-        multiplier: 1
-      },
       "google/gemma-4-26b-a4b-it": {
         input: 7e-8,
         output: 34e-8,
@@ -45465,6 +45490,11 @@ var init_model_pricing = __esm({
         output: 34e-8,
         multiplier: 1
       },
+      "google/gemma-4-31b-it:batch": {
+        input: 39e-8,
+        output: 97e-8,
+        multiplier: 2
+      },
       "google/lyria-3-clip-preview": {
         input: 0,
         output: 0,
@@ -45517,6 +45547,11 @@ var init_model_pricing = __esm({
         output: 9e-7,
         multiplier: 2
       },
+      "mistralai/devstral-2512": {
+        input: 4e-7,
+        output: 2e-6,
+        multiplier: 3
+      },
       "mistralai/ministral-14b-2512": {
         input: 2e-7,
         output: 2e-7,
@@ -45557,6 +45592,11 @@ var init_model_pricing = __esm({
         output: 75e-7,
         multiplier: 12
       },
+      "mistralai/mistral-medium-3-5:batch": {
+        input: 75e-8,
+        output: 375e-8,
+        multiplier: 6
+      },
       "mistralai/mistral-medium-3.1": {
         input: 4e-7,
         output: 2e-6,
@@ -45628,16 +45668,16 @@ var init_model_pricing = __esm({
         output: 34e-7,
         multiplier: 5
       },
-      "moonshotai/kimi-k2.7-code:batch": {
-        input: 475e-9,
-        output: 2e-6,
-        multiplier: 3
-      },
       "moonshotai/kimi-k3": {
         input: 285e-8,
         output: 1425e-8,
         multiplier: 23
       },
+      "moonshotai/kimi-k3:batch": {
+        input: 3e-6,
+        output: 15e-6,
+        multiplier: 24
+      },
       // Nvidia Models
       "nvidia/nemotron-3-nano-30b-a3b": {
         input: 5e-8,
@@ -45654,10 +45694,10 @@ var init_model_pricing = __esm({
         output: 22e-7,
         multiplier: 4
       },
-      "nvidia/nemotron-3-ultra-550b-a55b:batch": {
-        input: 3e-7,
-        output: 18e-7,
-        multiplier: 3
+      "nvidia/nemotron-3.5-content-safety": {
+        input: 2e-7,
+        output: 2e-7,
+        multiplier: 1
       },
       "nvidia/nemotron-3.5-lightning": {
         input: 8e-8,
@@ -45785,11 +45825,6 @@ var init_model_pricing = __esm({
         output: 1e-5,
         multiplier: 15
       },
-      "openai/gpt-5-codex:batch": {
-        input: 625e-9,
-        output: 5e-6,
-        multiplier: 8
-      },
       "openai/gpt-5-image": {
         input: 1e-5,
         output: 1e-5,
@@ -45981,24 +46016,24 @@ var init_model_pricing = __esm({
         multiplier: 1
       },
       "openai/gpt-5.6-sol": {
-        input: 25e-7,
-        output: 15e-6,
-        multiplier: 23
+        input: 2e-6,
+        output: 1e-5,
+        multiplier: 16
       },
       "openai/gpt-5.6-sol-pro": {
-        input: 25e-7,
-        output: 15e-6,
-        multiplier: 23
+        input: 2e-6,
+        output: 1e-5,
+        multiplier: 16
       },
       "openai/gpt-5.6-sol-pro:batch": {
-        input: 125e-8,
-        output: 75e-7,
-        multiplier: 12
+        input: 1e-6,
+        output: 5e-6,
+        multiplier: 8
       },
       "openai/gpt-5.6-sol:batch": {
-        input: 125e-8,
-        output: 75e-7,
-        multiplier: 12
+        input: 1e-6,
+        output: 5e-6,
+        multiplier: 8
       },
       "openai/gpt-5.6-terra": {
         input: 2e-6,
@@ -46020,6 +46055,26 @@ var init_model_pricing = __esm({
         output: 6e-6,
         multiplier: 9
       },
+      "openai/gpt-6-astra": {
+        input: 1e-5,
+        output: 5e-5,
+        multiplier: 80
+      },
+      "openai/gpt-6-astra-pro": {
+        input: 1e-5,
+        output: 5e-5,
+        multiplier: 80
+      },
+      "openai/gpt-6-astra-pro:batch": {
+        input: 5e-6,
+        output: 25e-6,
+        multiplier: 40
+      },
+      "openai/gpt-6-astra:batch": {
+        input: 5e-6,
+        output: 25e-6,
+        multiplier: 40
+      },
       "openai/gpt-audio": {
         input: 25e-7,
         output: 1e-5,
@@ -46040,11 +46095,21 @@ var init_model_pricing = __esm({
         output: 17e-8,
         multiplier: 1
       },
+      "openai/gpt-oss-120b:batch": {
+        input: 15e-8,
+        output: 6e-7,
+        multiplier: 1
+      },
       "openai/gpt-oss-20b": {
         input: 3e-8,
         output: 14e-8,
         multiplier: 1
       },
+      "openai/gpt-oss-20b:batch": {
+        input: 5e-8,
+        output: 2e-7,
+        multiplier: 1
+      },
       "openai/gpt-oss-safeguard-20b": {
         input: 75e-9,
         output: 3e-7,
@@ -46060,16 +46125,6 @@ var init_model_pricing = __esm({
         output: 6e-4,
         multiplier: 1e3
       },
-      "openai/o1-pro:batch": {
-        input: 75e-6,
-        output: 3e-4,
-        multiplier: 500
-      },
-      "openai/o1:batch": {
-        input: 75e-7,
-        output: 3e-5,
-        multiplier: 50
-      },
       "openai/o3": {
         input: 2e-6,
         output: 8e-6,
@@ -46085,11 +46140,6 @@ var init_model_pricing = __esm({
         output: 44e-7,
         multiplier: 7
       },
-      "openai/o3-mini-high:batch": {
-        input: 55e-8,
-        output: 22e-7,
-        multiplier: 4
-      },
       "openai/o3-mini:batch": {
         input: 55e-8,
         output: 22e-7,
@@ -46100,11 +46150,6 @@ var init_model_pricing = __esm({
         output: 8e-5,
         multiplier: 133
       },
-      "openai/o3-pro:batch": {
-        input: 1e-5,
-        output: 4e-5,
-        multiplier: 67
-      },
       "openai/o3:batch": {
         input: 1e-6,
         output: 4e-6,
@@ -46120,11 +46165,6 @@ var init_model_pricing = __esm({
         output: 44e-7,
         multiplier: 7
       },
-      "openai/o4-mini-high:batch": {
-        input: 55e-8,
-        output: 22e-7,
-        multiplier: 4
-      },
       "openai/o4-mini:batch": {
         input: 55e-8,
         output: 22e-7,
@@ -46136,11 +46176,6 @@ var init_model_pricing = __esm({
         output: 4e-7,
         multiplier: 1
       },
-      "qwen/qwen-2.5-7b-instruct": {
-        input: 3e-7,
-        output: 3e-7,
-        multiplier: 1
-      },
       "qwen/qwen-plus": {
         input: 26e-8,
         output: 78e-8,
@@ -46151,15 +46186,10 @@ var init_model_pricing = __esm({
         output: 78e-8,
         multiplier: 1
       },
-      "qwen/qwen-plus-2025-07-28:thinking": {
-        input: 26e-8,
-        output: 78e-8,
-        multiplier: 1
-      },
       "qwen/qwen2.5-vl-72b-instruct": {
-        input: 25e-8,
-        output: 75e-8,
-        multiplier: 1
+        input: 8e-7,
+        output: 1e-6,
+        multiplier: 2
       },
       "qwen/qwen3-14b": {
         input: 12e-8,
@@ -46311,6 +46341,11 @@ var init_model_pricing = __esm({
         output: 15e-8,
         multiplier: 1
       },
+      "qwen/qwen3.5-9b:batch": {
+        input: 17e-8,
+        output: 25e-8,
+        multiplier: 1
+      },
       "qwen/qwen3.5-flash-02-23": {
         input: 65e-9,
         output: 26e-8,
@@ -46371,20 +46406,35 @@ var init_model_pricing = __esm({
         output: 6e-6,
         multiplier: 11
       },
-      "qwen/qwen3.8-max": {
+      "qwen/qwen3.8-2.4t-a95b:batch": {
+        input: 2e-6,
+        output: 6e-6,
+        multiplier: 11
+      },
+      "qwen/qwen3.8-27b": {
+        input: 24e-8,
+        output: 22e-7,
+        multiplier: 3
+      },
+      "qwen/qwen3.8-flash": {
+        input: 15e-8,
+        output: 47e-8,
+        multiplier: 1
+      },
+      "qwen/qwen3.8-max-0902": {
         input: 2e-6,
         output: 6e-6,
         multiplier: 11
       },
       // Xiaomi Models
       "xiaomi/mimo-v2.5": {
-        input: 14e-8,
-        output: 28e-8,
+        input: 13e-8,
+        output: 65e-8,
         multiplier: 1
       },
       "xiaomi/mimo-v2.5-pro": {
-        input: 435e-9,
-        output: 87e-8,
+        input: 39e-8,
+        output: 117e-8,
         multiplier: 2
       },
       // Default pricing (GPT-4o-mini baseline)
@@ -46403,6 +46453,259 @@ var init_model_registry_generated = __esm({
   "src/lib/model-registry-generated.ts"() {
     "use strict";
     GENERATED_MODELS = [
+      {
+        id: "openai/gpt-6-astra",
+        displayName: "GPT-6 Astra",
+        vendor: "openai",
+        category: "specialized",
+        capabilities: {
+          vision: true,
+          reasoning: true,
+          maxTokens: 128e3,
+          contextLength: 105e4
+        },
+        pricing: {
+          inputPer1M: 10,
+          outputPer1M: 50,
+          multiplier: 80
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788552838
+      },
+      {
+        id: "openai/gpt-6-astra:batch",
+        displayName: "GPT-6 Astra (batch)",
+        vendor: "openai",
+        category: "powerful",
+        capabilities: {
+          vision: true,
+          reasoning: true,
+          maxTokens: 128e3,
+          contextLength: 105e4
+        },
+        pricing: {
+          inputPer1M: 5,
+          outputPer1M: 25,
+          multiplier: 40
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788552838
+      },
+      {
+        id: "openai/gpt-6-astra-pro",
+        displayName: "GPT-6 Astra Pro",
+        vendor: "openai",
+        category: "specialized",
+        capabilities: {
+          vision: true,
+          reasoning: true,
+          maxTokens: 128e3,
+          contextLength: 105e4
+        },
+        pricing: {
+          inputPer1M: 10,
+          outputPer1M: 50,
+          multiplier: 80
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788552835
+      },
+      {
+        id: "openai/gpt-6-astra-pro:batch",
+        displayName: "GPT-6 Astra Pro (batch)",
+        vendor: "openai",
+        category: "powerful",
+        capabilities: {
+          vision: true,
+          reasoning: true,
+          maxTokens: 128e3,
+          contextLength: 105e4
+        },
+        pricing: {
+          inputPer1M: 5,
+          outputPer1M: 25,
+          multiplier: 40
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788552835
+      },
+      {
+        id: "qwen/qwen3.8-max-0902",
+        displayName: "Qwen3.8 Max (0902)",
+        vendor: "qwen",
+        category: "balanced",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 131072,
+          contextLength: 1e6
+        },
+        pricing: {
+          inputPer1M: 2,
+          outputPer1M: 6,
+          multiplier: 11
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788469704
+      },
+      {
+        id: "google/gemini-3.8-flash",
+        displayName: "Gemini 3.8 Flash",
+        vendor: "google",
+        category: "balanced",
+        capabilities: {
+          vision: true,
+          reasoning: true,
+          maxTokens: 65536,
+          contextLength: 1048576
+        },
+        pricing: {
+          inputPer1M: 0.75,
+          outputPer1M: 3.75,
+          multiplier: 6
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788362056
+      },
+      {
+        id: "google/gemini-3.8-flash:batch",
+        displayName: "Gemini 3.8 Flash (batch)",
+        vendor: "google",
+        category: "fast",
+        capabilities: {
+          vision: true,
+          reasoning: true,
+          maxTokens: 65536,
+          contextLength: 1048576
+        },
+        pricing: {
+          inputPer1M: 0.375,
+          outputPer1M: 1.875,
+          multiplier: 3
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788362056
+      },
+      {
+        id: "anthropic/claude-fable-5.1",
+        displayName: "Claude Fable 5.1",
+        vendor: "anthropic",
+        category: "specialized",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 128e3,
+          contextLength: 1e6
+        },
+        pricing: {
+          inputPer1M: 10,
+          outputPer1M: 50,
+          multiplier: 80
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1788285838
+      },
+      {
+        id: "qwen/qwen3.8-flash",
+        displayName: "Qwen3.8 Flash",
+        vendor: "qwen",
+        category: "fast",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 131072,
+          contextLength: 1e6
+        },
+        pricing: {
+          inputPer1M: 0.15,
+          outputPer1M: 0.47,
+          multiplier: 1
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1787773060
+      },
+      {
+        id: "deepseek/deepseek-v4-flash-vision-exp",
+        displayName: "DeepSeek V4 Flash Vision Exp",
+        vendor: "deepseek",
+        category: "fast",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 262144,
+          contextLength: 1048576
+        },
+        pricing: {
+          inputPer1M: 0.2156,
+          outputPer1M: 0.6468,
+          multiplier: 1
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1787311563
+      },
+      {
+        id: "qwen/qwen3.8-27b",
+        displayName: "Qwen3.8 27B",
+        vendor: "qwen",
+        category: "fast",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 235929,
+          contextLength: 1e6
+        },
+        pricing: {
+          inputPer1M: 0.24,
+          outputPer1M: 2.2,
+          multiplier: 3
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1786722910
+      },
       {
         id: "google/gemini-3.7-flash",
         displayName: "Gemini 3.7 Flash",
@@ -46438,9 +46741,9 @@ var init_model_registry_generated = __esm({
           contextLength: 1048576
         },
         pricing: {
-          inputPer1M: 0.1875,
-          outputPer1M: 0.9375,
-          multiplier: 2
+          inputPer1M: 0.375,
+          outputPer1M: 1.875,
+          multiplier: 3
         },
         availability: {
           mainChat: false,
@@ -46472,17 +46775,63 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1786551702
       },
+      {
+        id: "qwen/qwen3.8-2.4t-a95b:batch",
+        displayName: "Qwen3.8 2.4T A95B (batch)",
+        vendor: "qwen",
+        category: "balanced",
+        capabilities: {
+          vision: false,
+          reasoning: false,
+          maxTokens: 909e3,
+          contextLength: 101e4
+        },
+        pricing: {
+          inputPer1M: 2,
+          outputPer1M: 6,
+          multiplier: 11
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1786551702
+      },
       {
         id: "deepseek/deepseek-v4-pro-0813",
         displayName: "DeepSeek V4 Pro 0813",
         vendor: "deepseek",
-        category: "balanced",
+        category: "fast",
         capabilities: {
           vision: false,
           reasoning: true,
           maxTokens: 393216,
           contextLength: 1048576
         },
+        pricing: {
+          inputPer1M: 0.5808,
+          outputPer1M: 1.7424,
+          multiplier: 3
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1786549364
+      },
+      {
+        id: "deepseek/deepseek-v4-pro-0813:batch",
+        displayName: "DeepSeek V4 Pro 0813 (batch)",
+        vendor: "deepseek",
+        category: "balanced",
+        capabilities: {
+          vision: false,
+          reasoning: true,
+          maxTokens: 943718,
+          contextLength: 1048576
+        },
         pricing: {
           inputPer1M: 1.32,
           outputPer1M: 3.96,
@@ -46519,42 +46868,42 @@ var init_model_registry_generated = __esm({
         createdAt: 1786452751
       },
       {
-        id: "qwen/qwen3.8-max",
-        displayName: "Qwen3.8 Max",
-        vendor: "qwen",
-        category: "balanced",
+        id: "deepseek/deepseek-v4-flash-0731",
+        displayName: "DeepSeek V4 Flash 0731",
+        vendor: "deepseek",
+        category: "fast",
         capabilities: {
-          vision: true,
+          vision: false,
           reasoning: false,
-          maxTokens: 131072,
-          contextLength: 1e6
+          maxTokens: 384e3,
+          contextLength: 1048576
         },
         pricing: {
-          inputPer1M: 2,
-          outputPer1M: 6,
-          multiplier: 11
+          inputPer1M: 0.06,
+          outputPer1M: 0.18,
+          multiplier: 1
         },
         availability: {
           mainChat: false,
           experts: true,
           embeds: true
         },
-        createdAt: 1785731612
+        createdAt: 1785478908
       },
       {
-        id: "deepseek/deepseek-v4-flash-0731",
-        displayName: "DeepSeek V4 Flash 0731",
+        id: "deepseek/deepseek-v4-flash-0731:batch",
+        displayName: "DeepSeek V4 Flash 0731 (batch)",
         vendor: "deepseek",
         category: "fast",
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 384e3,
+          maxTokens: 943718,
           contextLength: 1048576
         },
         pricing: {
-          inputPer1M: 0.08,
-          outputPer1M: 0.18,
+          inputPer1M: 0.14,
+          outputPer1M: 0.28,
           multiplier: 1
         },
         availability: {
@@ -46725,6 +47074,29 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1784215858
       },
+      {
+        id: "moonshotai/kimi-k3:batch",
+        displayName: "MoonshotAI: Kimi K3 (batch)",
+        vendor: "moonshotai",
+        category: "powerful",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 943718,
+          contextLength: 1048576
+        },
+        pricing: {
+          inputPer1M: 3,
+          outputPer1M: 15,
+          multiplier: 24
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1784215858
+      },
       {
         id: "openai/gpt-5.6-luna-pro",
         displayName: "GPT-5.6 Luna Pro",
@@ -46913,7 +47285,7 @@ var init_model_registry_generated = __esm({
         id: "openai/gpt-5.6-sol-pro",
         displayName: "GPT-5.6 Sol Pro",
         vendor: "openai",
-        category: "powerful",
+        category: "balanced",
         capabilities: {
           vision: true,
           reasoning: true,
@@ -46921,9 +47293,9 @@ var init_model_registry_generated = __esm({
           contextLength: 105e4
         },
         pricing: {
-          inputPer1M: 2.5,
-          outputPer1M: 15,
-          multiplier: 23
+          inputPer1M: 2,
+          outputPer1M: 10,
+          multiplier: 16
         },
         availability: {
           mainChat: false,
@@ -46944,9 +47316,9 @@ var init_model_registry_generated = __esm({
           contextLength: 105e4
         },
         pricing: {
-          inputPer1M: 1.25,
-          outputPer1M: 7.5,
-          multiplier: 12
+          inputPer1M: 1,
+          outputPer1M: 5,
+          multiplier: 8
         },
         availability: {
           mainChat: false,
@@ -46959,7 +47331,7 @@ var init_model_registry_generated = __esm({
         id: "openai/gpt-5.6-sol",
         displayName: "GPT-5.6 Sol",
         vendor: "openai",
-        category: "powerful",
+        category: "balanced",
         capabilities: {
           vision: true,
           reasoning: true,
@@ -46967,9 +47339,9 @@ var init_model_registry_generated = __esm({
           contextLength: 105e4
         },
         pricing: {
-          inputPer1M: 2.5,
-          outputPer1M: 15,
-          multiplier: 23
+          inputPer1M: 2,
+          outputPer1M: 10,
+          multiplier: 16
         },
         availability: {
           mainChat: false,
@@ -46990,9 +47362,9 @@ var init_model_registry_generated = __esm({
           contextLength: 105e4
         },
         pricing: {
-          inputPer1M: 1.25,
-          outputPer1M: 7.5,
-          multiplier: 12
+          inputPer1M: 1,
+          outputPer1M: 5,
+          multiplier: 8
         },
         availability: {
           mainChat: false,
@@ -47070,29 +47442,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1781266361
       },
-      {
-        id: "moonshotai/kimi-k2.7-code:batch",
-        displayName: "MoonshotAI: Kimi K2.7 Code (batch)",
-        vendor: "moonshotai",
-        category: "fast",
-        capabilities: {
-          vision: true,
-          reasoning: false,
-          maxTokens: 4e4,
-          contextLength: 262144
-        },
-        pricing: {
-          inputPer1M: 0.475,
-          outputPer1M: 2,
-          multiplier: 3
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1781266361
-      },
       {
         id: "anthropic/claude-fable-5",
         displayName: "Claude Fable 5",
@@ -47125,7 +47474,7 @@ var init_model_registry_generated = __esm({
           vision: false,
           reasoning: false,
           maxTokens: 16384,
-          contextLength: 512288
+          contextLength: 262144
         },
         pricing: {
           inputPer1M: 0.5,
@@ -47139,29 +47488,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1780551208
       },
-      {
-        id: "nvidia/nemotron-3-ultra-550b-a55b:batch",
-        displayName: "NVIDIA: Nemotron 3 Ultra (batch)",
-        vendor: "nvidia",
-        category: "fast",
-        capabilities: {
-          vision: false,
-          reasoning: false,
-          maxTokens: 4e4,
-          contextLength: 512288
-        },
-        pricing: {
-          inputPer1M: 0.3,
-          outputPer1M: 1.8,
-          multiplier: 3
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1780551208
-      },
       {
         id: "qwen/qwen3.7-plus",
         displayName: "Qwen3.7 Plus",
@@ -47354,7 +47680,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 209715,
           contextLength: 262144
         },
         pricing: {
@@ -47369,6 +47695,29 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1777570439
       },
+      {
+        id: "mistralai/mistral-medium-3-5:batch",
+        displayName: "Mistral Medium 3.5 (batch)",
+        vendor: "mistralai",
+        category: "balanced",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 26214,
+          contextLength: 32768
+        },
+        pricing: {
+          inputPer1M: 0.75,
+          outputPer1M: 3.75,
+          multiplier: 6
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1777570439
+      },
       {
         id: "qwen/qwen3.5-plus-20260420",
         displayName: "Qwen3.5 Plus 2026-04-20",
@@ -47469,7 +47818,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 262144,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -47584,13 +47933,13 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: true,
-          maxTokens: 16384,
+          maxTokens: 943718,
           contextLength: 1048576
         },
         pricing: {
-          inputPer1M: 1.3,
-          outputPer1M: 2.6,
-          multiplier: 5
+          inputPer1M: 1.2,
+          outputPer1M: 1.2,
+          multiplier: 3
         },
         availability: {
           mainChat: false,
@@ -47634,8 +47983,8 @@ var init_model_registry_generated = __esm({
           contextLength: 1048576
         },
         pricing: {
-          inputPer1M: 0.435,
-          outputPer1M: 0.87,
+          inputPer1M: 0.39,
+          outputPer1M: 1.17,
           multiplier: 2
         },
         availability: {
@@ -47657,8 +48006,8 @@ var init_model_registry_generated = __esm({
           contextLength: 1048576
         },
         pricing: {
-          inputPer1M: 0.14,
-          outputPer1M: 0.28,
+          inputPer1M: 0.13,
+          outputPer1M: 0.65,
           multiplier: 1
         },
         availability: {
@@ -47676,7 +48025,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 262144,
+          maxTokens: 16384,
           contextLength: 262144
         },
         pricing: {
@@ -47760,6 +48109,29 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1775148486
       },
+      {
+        id: "google/gemma-4-31b-it:batch",
+        displayName: "Gemma 4 31B (batch)",
+        vendor: "google",
+        category: "fast",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 235929,
+          contextLength: 262144
+        },
+        pricing: {
+          inputPer1M: 0.39,
+          outputPer1M: 0.97,
+          multiplier: 2
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1775148486
+      },
       {
         id: "qwen/qwen3.6-plus",
         displayName: "Qwen3.6 Plus",
@@ -47883,7 +48255,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 209715,
           contextLength: 262144
         },
         pricing: {
@@ -47929,7 +48301,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 262144,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -47944,6 +48316,29 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1773152396
       },
+      {
+        id: "qwen/qwen3.5-9b:batch",
+        displayName: "Qwen3.5-9B (batch)",
+        vendor: "qwen",
+        category: "fast",
+        capabilities: {
+          vision: true,
+          reasoning: false,
+          maxTokens: 235929,
+          contextLength: 262144
+        },
+        pricing: {
+          inputPer1M: 0.17,
+          outputPer1M: 0.25,
+          multiplier: 1
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1773152396
+      },
       {
         id: "openai/gpt-5.4-pro",
         displayName: "GPT-5.4 Pro",
@@ -48113,7 +48508,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 262144,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -48366,7 +48761,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 262144,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -48389,7 +48784,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 64e3,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -48527,7 +48922,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 228e3,
+          maxTokens: 32768,
           contextLength: 262144
         },
         pricing: {
@@ -48657,6 +49052,29 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1765389775
       },
+      {
+        id: "mistralai/devstral-2512",
+        displayName: "Devstral 2 2512",
+        vendor: "mistralai",
+        category: "fast",
+        capabilities: {
+          vision: false,
+          reasoning: false,
+          maxTokens: 209715,
+          contextLength: 262144
+        },
+        pricing: {
+          inputPer1M: 0.4,
+          outputPer1M: 2,
+          multiplier: 3
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1765285419
+      },
       {
         id: "openai/gpt-5.1-codex-max",
         displayName: "GPT-5.1-Codex-Max",
@@ -48711,7 +49129,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 209715,
           contextLength: 262144
         },
         pricing: {
@@ -48734,7 +49152,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 209715,
           contextLength: 262144
         },
         pricing: {
@@ -48757,7 +49175,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 104857,
           contextLength: 131072
         },
         pricing: {
@@ -48780,7 +49198,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 209715,
           contextLength: 262144
         },
         pricing: {
@@ -48803,7 +49221,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 163840,
+          maxTokens: 147456,
           contextLength: 163840
         },
         pricing: {
@@ -48941,7 +49359,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: true,
-          maxTokens: 100352,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -48987,8 +49405,8 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4e4,
-          contextLength: 32e3
+          maxTokens: 26214,
+          contextLength: 32768
         },
         pricing: {
           inputPer1M: 0.1,
@@ -49240,7 +49658,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 65536,
+          maxTokens: 147456,
           contextLength: 163840
         },
         pricing: {
@@ -49347,29 +49765,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1758662707
       },
-      {
-        id: "openai/gpt-5-codex:batch",
-        displayName: "GPT-5 Codex (batch)",
-        vendor: "openai",
-        category: "balanced",
-        capabilities: {
-          vision: true,
-          reasoning: true,
-          maxTokens: 128e3,
-          contextLength: 4e5
-        },
-        pricing: {
-          inputPer1M: 0.625,
-          outputPer1M: 5,
-          multiplier: 8
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1758643403
-      },
       {
         id: "deepseek/deepseek-v3.1-terminus",
         displayName: "DeepSeek V3.1 Terminus",
@@ -49378,7 +49773,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 163840,
+          maxTokens: 147456,
           contextLength: 163840
         },
         pricing: {
@@ -49424,7 +49819,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: true,
-          maxTokens: 262144,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -49485,29 +49880,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1757347599
       },
-      {
-        id: "qwen/qwen-plus-2025-07-28:thinking",
-        displayName: "Qwen Plus 0728 (thinking)",
-        vendor: "qwen",
-        category: "fast",
-        capabilities: {
-          vision: false,
-          reasoning: true,
-          maxTokens: 32768,
-          contextLength: 1e6
-        },
-        pricing: {
-          inputPer1M: 0.26,
-          outputPer1M: 0.78,
-          multiplier: 1
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1757347599
-      },
       {
         id: "moonshotai/kimi-k2-0905",
         displayName: "MoonshotAI: Kimi K2 0905",
@@ -49585,7 +49957,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 104857,
           contextLength: 131072
         },
         pricing: {
@@ -49654,7 +50026,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 128e3,
+          maxTokens: 36e4,
           contextLength: 4e5
         },
         pricing: {
@@ -49700,7 +50072,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 128e3,
+          maxTokens: 36e4,
           contextLength: 4e5
         },
         pricing: {
@@ -49746,7 +50118,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 131072,
+          maxTokens: 117964,
           contextLength: 131072
         },
         pricing: {
@@ -49761,6 +50133,29 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1754414231
       },
+      {
+        id: "openai/gpt-oss-120b:batch",
+        displayName: "gpt-oss-120b (batch)",
+        vendor: "openai",
+        category: "fast",
+        capabilities: {
+          vision: false,
+          reasoning: false,
+          maxTokens: 117964,
+          contextLength: 131072
+        },
+        pricing: {
+          inputPer1M: 0.15,
+          outputPer1M: 0.6,
+          multiplier: 1
+        },
+        availability: {
+          mainChat: false,
+          experts: true,
+          embeds: true
+        },
+        createdAt: 1754414231
+      },
       {
         id: "openai/gpt-oss-20b",
         displayName: "gpt-oss-20b",
@@ -49769,7 +50164,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 131072,
+          maxTokens: 117964,
           contextLength: 131072
         },
         pricing: {
@@ -49815,7 +50210,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 204800,
           contextLength: 256e3
         },
         pricing: {
@@ -49861,7 +50256,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 262144,
+          maxTokens: 235929,
           contextLength: 262144
         },
         pricing: {
@@ -49884,8 +50279,8 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: true,
-          maxTokens: 32768,
-          contextLength: 262144
+          maxTokens: 117964,
+          contextLength: 131072
         },
         pricing: {
           inputPer1M: 0.23,
@@ -50152,29 +50547,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1749598352
       },
-      {
-        id: "openai/o3-pro:batch",
-        displayName: "o3 Pro (batch)",
-        vendor: "openai",
-        category: "specialized",
-        capabilities: {
-          vision: true,
-          reasoning: true,
-          maxTokens: 1e5,
-          contextLength: 2e5
-        },
-        pricing: {
-          inputPer1M: 10,
-          outputPer1M: 40,
-          multiplier: 67
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1749598352
-      },
       {
         id: "google/gemini-2.5-pro-preview",
         displayName: "Gemini 2.5 Pro Preview 06-05",
@@ -50275,7 +50647,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 104857,
           contextLength: 131072
         },
         pricing: {
@@ -50451,29 +50823,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1744824212
       },
-      {
-        id: "openai/o4-mini-high:batch",
-        displayName: "o4 Mini High (batch)",
-        vendor: "openai",
-        category: "fast",
-        capabilities: {
-          vision: true,
-          reasoning: true,
-          maxTokens: 1e5,
-          contextLength: 2e5
-        },
-        pricing: {
-          inputPer1M: 0.55,
-          outputPer1M: 2.2,
-          multiplier: 4
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1744824212
-      },
       {
         id: "openai/o3",
         displayName: "o3",
@@ -50574,7 +50923,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 942818,
           contextLength: 1047576
         },
         pricing: {
@@ -50620,7 +50969,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 32768,
+          maxTokens: 942818,
           contextLength: 1047576
         },
         pricing: {
@@ -50666,7 +51015,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: true,
           reasoning: false,
-          maxTokens: 32768,
+          maxTokens: 942818,
           contextLength: 1047576
         },
         pricing: {
@@ -50827,7 +51176,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 26214,
           contextLength: 32768
         },
         pricing: {
@@ -50865,29 +51214,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1739372611
       },
-      {
-        id: "openai/o3-mini-high:batch",
-        displayName: "o3 Mini High (batch)",
-        vendor: "openai",
-        category: "fast",
-        capabilities: {
-          vision: false,
-          reasoning: true,
-          maxTokens: 1e5,
-          contextLength: 2e5
-        },
-        pricing: {
-          inputPer1M: 0.55,
-          outputPer1M: 2.2,
-          multiplier: 4
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1739372611
-      },
       {
         id: "qwen/qwen-plus",
         displayName: "Qwen-Plus",
@@ -51026,29 +51352,6 @@ var init_model_registry_generated = __esm({
         },
         createdAt: 1734459999
       },
-      {
-        id: "openai/o1:batch",
-        displayName: "o1 (batch)",
-        vendor: "openai",
-        category: "powerful",
-        capabilities: {
-          vision: true,
-          reasoning: true,
-          maxTokens: 1e5,
-          contextLength: 2e5
-        },
-        pricing: {
-          inputPer1M: 7.5,
-          outputPer1M: 30,
-          multiplier: 50
-        },
-        availability: {
-          mainChat: false,
-          experts: true,
-          embeds: true
-        },
-        createdAt: 1734459999
-      },
       {
         id: "meta-llama/llama-3.3-70b-instruct",
         displayName: "Llama 3.3 70B Instruct",
@@ -51172,7 +51475,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 104857,
           contextLength: 131072
         },
         pricing: {
@@ -51425,7 +51728,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 52428,
           contextLength: 65536
         },
         pricing: {
@@ -51517,7 +51820,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4e4,
+          maxTokens: 102400,
           contextLength: 128e3
         },
         pricing: {
@@ -51540,7 +51843,7 @@ var init_model_registry_generated = __esm({
         capabilities: {
           vision: false,
           reasoning: false,
-          maxTokens: 4096,
+          maxTokens: 3685,
           contextLength: 4095
         },
         pricing: {
@@ -51693,7 +51996,7 @@ function resolveBedrockModelId(anthropicModelId, awsRegion) {
   return `${getBedrockRegionPrefix(awsRegion)}.anthropic.${anthropicModelId}-v1:0`;
 }
 function toModelDefinition(model) {
-  if (model.vendor === "anthropic") {
+  if (model.vendor === "anthropic" && !ANTHROPIC_OPENROUTER_ONLY.has(model.id)) {
     return {
       ...model,
       provider: "anthropic",
@@ -51732,7 +52035,7 @@ function calculateMaxOutputTokens(modelId, inputTokens) {
   const fits = contextLength - inputTokens - CONTEXT_BUFFER;
   return Math.max(0, Math.min(OUTPUT_TOKEN_CEILING, maxOutputTokens, fits));
 }
-var BEDROCK_MODEL_DEFS, GOOGLE_DIRECT_MODELS, MODEL_REGISTRY, CATEGORY_LABELS, DEFAULT_CONTEXT_LENGTH, DEFAULT_MAX_OUTPUT_TOKENS, OUTPUT_TOKEN_CEILING, CONTEXT_BUFFER;
+var BEDROCK_MODEL_DEFS, GOOGLE_DIRECT_MODELS, ANTHROPIC_OPENROUTER_ONLY, MODEL_REGISTRY, CATEGORY_LABELS, DEFAULT_CONTEXT_LENGTH, DEFAULT_MAX_OUTPUT_TOKENS, OUTPUT_TOKEN_CEILING, CONTEXT_BUFFER;
 var init_model_registry = __esm({
   "src/lib/model-registry.ts"() {
     "use strict";
@@ -51790,8 +52093,13 @@ var init_model_registry = __esm({
     GOOGLE_DIRECT_MODELS = {
       "google/gemini-3.1-pro-preview": "gemini-3.1-pro-preview",
       "google/gemini-3-pro-preview": "gemini-3-pro-preview",
-      "google/gemini-3.7-flash": "gemini-3.7-flash"
+      "google/gemini-3.7-flash": "gemini-3.7-flash",
+      "google/gemini-3.8-flash": "gemini-3.8-flash"
     };
+    ANTHROPIC_OPENROUTER_ONLY = /* @__PURE__ */ new Set([
+      "anthropic/claude-fable-5",
+      "anthropic/claude-fable-5.1"
+    ]);
     MODEL_REGISTRY = GENERATED_MODELS.map(toModelDefinition);
     CATEGORY_LABELS = {
       fast: "Fast (Budget)",
@@ -136786,7 +137094,7 @@ function normalizeClaudeModelId(modelId) {
 }
 function usesAdaptiveThinking(modelId) {
   const id = normalizeClaudeModelId(modelId);
-  return id.includes("claude-opus-4-7") || id.includes("claude-opus-4-8") || id.includes("claude-opus-5") || id.includes("claude-sonnet-5");
+  return id.includes("claude-opus-4-7") || id.includes("claude-opus-4-8") || id.includes("claude-opus-5") || id.includes("claude-sonnet-5") || id.includes("claude-fable-5");
 }
 function supportsSamplingParameters(modelId) {
   return !usesAdaptiveThinking(modelId);
@@ -137256,6 +137564,9 @@ function extractClaudeModel(modelId) {
     "anthropic/claude-opus-4.7": "claude-opus-4-7",
     "anthropic/claude-opus-4.8": "claude-opus-4-8",
     "anthropic/claude-opus-5": "claude-opus-5",
+    // Claude Fable models
+    "anthropic/claude-fable-5": "claude-fable-5",
+    "anthropic/claude-fable-5.1": "claude-fable-5-1",
     // Claude Haiku 4.5 models
     "anthropic/claude-haiku-4.5": "claude-haiku-4-5-20251001",
     "anthropic/claude-4.5-haiku": "claude-haiku-4-5-20251001",
@@ -164418,17 +164729,20 @@ var init_model_config = __esm({
       "anthropic/claude-4.5-sonnet": { vision: true, reasoning: true },
       "anthropic/claude-opus-4.7": { vision: true, reasoning: false },
       "anthropic/claude-opus-5": { vision: true, reasoning: true },
+      "anthropic/claude-fable-5.1": { vision: true, reasoning: true },
       "google/gemini-2.5-flash": { vision: true, reasoning: true },
       "google/gemini-2.5-flash-lite": { vision: true, reasoning: true },
       "google/gemini-2.5-pro": { vision: true, reasoning: true },
       "google/gemini-3.1-pro-preview": { vision: true, reasoning: true },
       "google/gemini-3-pro-preview": { vision: true, reasoning: true },
       "google/gemini-3.7-flash": { vision: true, reasoning: true },
+      "google/gemini-3.8-flash": { vision: true, reasoning: true },
       "openai/gpt-4o": { vision: true, reasoning: false },
       "openai/gpt-5-mini": { vision: true, reasoning: false },
       "openai/gpt-5.2-codex": { vision: true, reasoning: true },
       "openai/gpt-5.5": { vision: true, reasoning: true },
       "openai/gpt-5.6-sol": { vision: true, reasoning: true },
+      "openai/gpt-6-astra": { vision: true, reasoning: true },
       "deepseek/deepseek-r1": { vision: false, reasoning: true }
     };
   }
@@ -165611,8 +165925,8 @@ var init_model_tiers = __esm({
             zdr: true
           },
           {
-            id: "google/gemini-3.7-flash",
-            displayName: "Gemini 3.7 Flash",
+            id: "google/gemini-3.8-flash",
+            displayName: "Gemini 3.8 Flash",
             description: "Fast coding and agentic workflows, 1M context, native multimodal",
             provider: "Google",
             multiLingual: true,
@@ -165673,11 +165987,24 @@ var init_model_tiers = __esm({
             zdr: true
           },
           {
-            id: "openai/gpt-5.6-sol",
-            displayName: "GPT-5.6 Sol",
+            id: "anthropic/claude-fable-5.1",
+            displayName: "Claude Fable 5.1",
+            description: "Long-running agentic coding, large refactors, and front-end work",
+            provider: "Anthropic",
+            costTier: "Very Expensive",
+            // Not a mistake and not safe to flip back: Fable 5.1 has no zero-data-retention
+            // endpoint on any route we can reach. Bedrock answers `data retention mode
+            // 'default' is not available for this model`, and OpenRouter answers
+            // `No endpoints found matching your data policy (Zero data retention)`.
+            // Forcing ZDR makes every request fail.
+            zdr: false
+          },
+          {
+            id: "openai/gpt-6-astra",
+            displayName: "GPT-6 Astra",
             description: "1M+ context, frontier reasoning, multimodal, and agentic tool use",
             provider: "OpenAI",
-            costTier: "Expensive",
+            costTier: "Very Expensive",
             multiLingual: true,
             zdr: true,
             effort: "medium"
@@ -217138,7 +217465,9 @@ var init_spawn_subagent = __esm({
       allowedTools: external_exports.array(external_exports.string()).describe(
         'RAW MCP tool names the subagent may use, e.g. "agentic-search-fragments" \u2014 NOT prefixed forms like "mcp__usable__agentic-search-fragments". `spawn_subagent`, `await_subagents`, and `list_subagents` are forbidden in this list (no recursion in v1).'
       ),
-      systemPrompt: external_exports.string().max(8e3).optional().describe("Optional system prompt override. If omitted, a generic worker-agent system prompt is used.")
+      systemPrompt: external_exports.string().max(8e3).optional().describe(
+        "Optional system prompt override. If omitted, a generic worker-agent system prompt is used."
+      )
     });
     DEFAULT_SYSTEM_PROMPT = `You are a focused worker subagent spawned by a main orchestrator agent.
 
@@ -217184,7 +217513,7 @@ HOW TO USE THIS:
 - \`prompt\`: the actual work task. The subagent receives this as its user message.
 - \`model\`: **default to \`quick-thinking\` (alias for \`anthropic/claude-haiku-4.5\`) unless you have a specific reason not to.** Subagents are billed in addition to your own turn \u2014 every Sonnet-class subagent doubles cost. Guidance:
     - **\`quick-thinking\` / \`anthropic/claude-haiku-4.5\`** (cheapest, fast, strong tool use) \u2014 use this for: web/Usable search, summarisation, classification, simple extraction, drafting, tabular work, anything where one-shot quality is "good enough". Should be your **default**.
-    - **\`google/gemini-3.7-flash\`** (cheap, huge context, multimodal/multi-lingual) \u2014 use this when: you need to read a 100k+-token document, the task is multi-lingual, or there's image/audio input.
+    - **\`google/gemini-3.8-flash\`** (cheap, huge context, multimodal/multi-lingual) \u2014 use this when: you need to read a 100k+-token document, the task is multi-lingual, or there's image/audio input.
     - **\`deep-thinking\` / \`anthropic/claude-sonnet-5\`** (5\u201310\xD7 the cost of Haiku) \u2014 only use when: the work requires multi-step reasoning over conflicting evidence, code generation against a specific architecture, or careful policy/risk judgement. If you find yourself reaching for this for "research", reach for Haiku first.
     - **\`google/gemini-3.1-pro-preview\`** \u2014 only when both massive context AND deep reasoning are required.
 - \`allowedTools\`: tool names exactly as they appear in YOUR own tool list. Raw MCP names (no \`mcp__usable__\` prefix) for built-in / MCP tools; \`parent_*\` prefixed names for embed-host tools registered by the surrounding application (e.g. \`parent_search_vessel_registry\`). Subagents in embed contexts CAN use parent tools \u2014 list them here just like any other tool.
@@ -217387,7 +217716,13 @@ HOW TO USE THIS:
                 tokenUsage: response.tokenUsage
               }
             });
-            const relayed = await maybeStartBackgroundRelay(redis, context, conversationId, taskId, success2);
+            const relayed = await maybeStartBackgroundRelay(
+              redis,
+              context,
+              conversationId,
+              taskId,
+              success2
+            );
             if (!relayed) {
               await maybeEmitPingPrompt(redis, conversationId, taskId);
             }
@@ -217405,7 +217740,13 @@ HOW TO USE THIS:
               taskId,
               data: { error: errorMessage }
             });
-            const relayed = await maybeStartBackgroundRelay(redis, context, conversationId, taskId, false);
+            const relayed = await maybeStartBackgroundRelay(
+              redis,
+              context,
+              conversationId,
+              taskId,
+              false
+            );
             if (!relayed) {
               await maybeEmitPingPrompt(redis, conversationId, taskId);
             }
@@ -292548,6 +292889,7 @@ function getStreamingMode() {
   return "sse";
 }
 function isClaudeModel(modelId) {
+  if (ANTHROPIC_OPENROUTER_ONLY.has(modelId)) return false;
   return modelId.startsWith("anthropic/") || modelId.includes("claude");
 }
 function filterImagesFromHistory(messages4, opts) {
@@ -297154,6 +297496,7 @@ var init_orchestrator = __esm({
     init_logger();
     init_model_config();
     init_model_tiers();
+    init_model_registry();
     init_mcp();
     init_model_registry();
     init_openrouter_direct();
@@ -316174,7 +316517,7 @@ init_tui_select();
 init_model_registry();
 
 // package.json
-var version2 = "1.198.3";
+var version2 = "1.199.0";
 
 // src/adapters/cli/model-catalog.ts
 init_codex_auth();