modelmix 5.1.4 → 5.1.6

package/README.md

@@ -151,7 +151,7 @@ ModelMix provides convenient shorthand methods for quickly accessing different A
 | `gpt51()` | OpenAI | gpt-5.1 | [\$1.25][1] | [\$10.00][1] |
 | `gpt5mini()` | OpenAI | gpt-5-mini | [\$0.25][1] | [\$2.00][1] |
 | `gpt5nano()` | OpenAI | gpt-5-nano | [\$0.05][1] | [\$0.40][1] |
-| `gptOss()` | Together | gpt-oss-120B | [\$0.15][7] | [\$0.60][7] |
+| `gptOss()` | Cerebras/Groq/OpenRouter | gpt-oss-120B | [\$0.15][7] | [\$0.60][7] |
 | `fable5()` | Anthropic | claude-fable-5 | [\$10.00][2] | [\$50.00][2] |
 | `opus5()` | Anthropic | claude-opus-5 | [\$5.00][2] | [\$25.00][2] |
 | `opus48()` | Anthropic | claude-opus-4-8 | [\$5.00][2] | [\$25.00][2] |
@@ -171,8 +171,9 @@ ModelMix provides convenient shorthand methods for quickly accessing different A
 | `grok43()` | Grok | grok-4.3 | [\$1.25][6] | [\$2.50][6] |
 | `grok420multiAgent()` | Grok | grok-4.20-multi-agent-0309 | [\$1.25][6] | [\$2.50][6] |
 | `grok420()` | Grok | grok-4.20-0309 (†) | [\$1.25][6] | [\$2.50][6] |
+| `museGlimmer30b()` | Fireworks | models/muse-glimmer-30b | [\$0.35][17] | [\$1.50][17] |
 | `qwen35397b()` | OpenRouter | qwen/qwen3.5-397b-a17b | [\$0.385][14] | [\$2.45][14] |
-| `qwen36plus()` | Fireworks | qwen3p6-plus | [\$0.50][10] | [\$3.00][10] |
+| `qwen36plus()` | OpenRouter | qwen/qwen3.6-plus | [\$0.325][18] | [\$1.95][18] |
 | `qwen37plus()` | Fireworks | models/qwen3p7-plus | [\$0.40][10] | [\$1.60][10] |
 | `qwen38max()` | Fireworks | qwen3p8-2p4t-a95b | [\$2.00][10] | [\$6.00][10] |
 | `qwen3827b()` | OpenRouter | qwen/qwen3.8-27b | [\$0.45][15] | [\$3.20][15] |
@@ -180,7 +181,6 @@ ModelMix provides convenient shorthand methods for quickly accessing different A
 | `deepseekV4Pro()` | Fireworks | models/deepseek-v4-pro-0813 | [\$1.32][12] | [\$3.96][12] |
 | `GLM53()` | OpenRouter | z-ai/glm-5.3 | [\$1.40][16] | [\$4.40][16] |
 | `GLM52()` | Together | zai-org/GLM-5.2 | [\$1.40][7] | [\$4.40][7] |
-| `GLM51()` | Fireworks | models/glm-5p1 | [\$1.05][10] | [\$3.50][10] |
 | `minimaxM3()` | MiniMax | MiniMax-M3 | [\$0.30][9] | [\$1.20][9] |
 | `minimaxM27()` | MiniMax | MiniMax-M2.7 | [\$0.30][9] | [\$1.20][9] |
 | `sonar()` | Perplexity | sonar | [\$1.00][4] | [\$1.00][4] |
@@ -189,11 +189,16 @@ ModelMix provides convenient shorthand methods for quickly accessing different A
 | `hermes4405b()` | OpenRouter | nousresearch/hermes-4-405b | [\$1.00][13] | [\$3.00][13] |
 | `hermes3()` | Lambda | Hermes-3-Llama-3.1-405B-FP8 | [\$0.80][8] | [\$0.80][8] |
 | `kimiK3()` | Moonshot | kimi-k3 | [\$3.00][11] | [\$15.00][11] |
+| `kimiK27Code()` | Together | Kimi-K2.7-Code | [\$0.95][7] | [\$4.00][7] |
 | `kimiK25()` | Together | Kimi-K2.5 | [\$0.50][7] | [\$2.80][7] |
 | `kimiK26()` | Fireworks | models/kimi-k2p6 | [\$0.95][10] | [\$4.00][10] |
 
 Gemini 3.7 Flash and 3.6 Flash use Google's introductory standard pricing through December 31, 2026; standard rates double on January 1, 2027.
 
+`museGlimmer30b()` uses Fireworks first and OpenRouter as its default fallback. NVIDIA NIM and Together are also available through `mix.nvidia` and `mix.together`; disable either default provider with `mix.fireworks: false` or `mix.openrouter: false`.
+
+The multi-provider shortcuts also expose the current catalog alternatives: `gptOss()` supports NVIDIA and Fireworks; `qwen37plus()` supports Together; `kimiK27Code()` supports Fireworks and OpenRouter; `kimiK3()` supports Fireworks, OpenRouter, and Together; `GLM52()` supports Fireworks and OpenRouter; and both MiniMax shortcuts support Fireworks. `minimaxM27()` keeps every enabled provider in its fallback chain. `qwen36plus()` now defaults to OpenRouter because Fireworks retired its serverless deployment; Fireworks remains available explicitly for private or on-demand deployments.
+
 [1]: https://platform.openai.com/docs/pricing "Pricing | OpenAI"
 [2]: https://docs.anthropic.com/en/docs/about-claude/pricing "Pricing - Anthropic"
 [3]: https://ai.google.dev/gemini-api/docs/pricing "Google AI for Developers"
@@ -210,6 +215,8 @@ Gemini 3.7 Flash and 3.6 Flash use Google's introductory standard pricing throug
 [14]: https://openrouter.ai/qwen/qwen3.5-397b-a17b "Qwen3.5 397B A17B on OpenRouter"
 [15]: https://openrouter.ai/qwen/qwen3.8-27b "Qwen3.8 27B on OpenRouter"
 [16]: https://openrouter.ai/z-ai/glm-5.3 "GLM 5.3 on OpenRouter"
+[17]: https://fireworks.ai/models/fireworks/muse-glimmer-30b "Muse Glimmer 30B on Fireworks"
+[18]: https://openrouter.ai/qwen/qwen3.6-plus "Qwen 3.6 Plus on OpenRouter"
 
 Each method accepts optional `options`, `config`, and (for multi-provider methods) `mix` parameters to customize behavior.  
 

package/demo/custom.js

@@ -50,8 +50,8 @@ class MixTogether extends MixCustom {
     }
 }
 
-mmix.attach('Qwen/Qwen3-235B-A22B-fp8-tput', new MixTogether());
+mmix.attach('Qwen/Qwen3.6-Plus', new MixTogether());
 
 let r = mmix.addText('hi there');
 r = await r.addText('do you like cats?').message();
-console.log(r);
+console.log(r);

package/demo/short.js

@@ -17,8 +17,9 @@ const mmix = await ModelMix.new(setup)
     .qwen35397b() // (fallback 5) OpenRouter qwen/qwen3.5-397b-a17b
     .qwen3827b() // (fallback 6) OpenRouter qwen/qwen3.8-27b
     .GLM53() // (fallback 7) OpenRouter z-ai/glm-5.3
-    .hermes470b() // (fallback 8) OpenRouter nousresearch/hermes-4-70b
-    .hermes4405b() // (fallback 9) OpenRouter nousresearch/hermes-4-405b
+    .museGlimmer30b({ mix: { fireworks: false, openrouter: true } }) // (fallback 8) OpenRouter meta/muse-glimmer-30b
+    .hermes470b() // (fallback 9) OpenRouter nousresearch/hermes-4-70b
+    .hermes4405b() // (fallback 10) OpenRouter nousresearch/hermes-4-405b
     .addText("What's your name?");
 
 console.log(await mmix.message());

package/effort.js

@@ -45,6 +45,9 @@ const OPENAI_MODEL_LEVELS = {
     'gpt-5.6-sol': ['none', 'low', 'medium', 'high', 'xhigh', 'max'],
     'gpt-5.6-terra': ['none', 'low', 'medium', 'high', 'xhigh', 'max'],
     'gpt-5.6-luna': ['none', 'low', 'medium', 'high', 'xhigh', 'max'],
+    'meta/muse-glimmer-30b': ['low', 'medium', 'high'],
+    'accounts/fireworks/models/muse-glimmer-30b': ['low', 'medium', 'high', 'xhigh'],
+    'meta-models/Muse-Glimmer-30B': ['low', 'medium', 'high', 'xhigh'],
     'accounts/fireworks/models/qwen3p8-2p4t-a95b': ['none', 'low', 'medium', 'high'],
     'qwen/qwen3.8-27b': ['low', 'medium', 'xhigh'],
     'z-ai/glm-5.3': ['low', 'high', 'max'],
@@ -55,6 +58,14 @@ const OPENAI_MODEL_LEVELS = {
     'gpt-5.3-codex': ['low', 'medium', 'high', 'xhigh'],
     'gpt-oss-120b': ['low', 'medium', 'high'],
     'openai/gpt-oss-120b': ['low', 'medium', 'high'],
+    'accounts/fireworks/models/gpt-oss-120b': ['low', 'medium', 'high'],
+    'kimi-k3': ['low', 'high', 'max'],
+    'moonshotai/kimi-k3': ['low', 'high', 'max'],
+    'accounts/fireworks/models/kimi-k3': ['low', 'high', 'max'],
+    'moonshotai/Kimi-K3': ['low', 'high', 'max'],
+    'zai-org/GLM-5.2': ['high', 'xhigh'],
+    'accounts/fireworks/models/glm-5p2': ['high', 'xhigh'],
+    'z-ai/glm-5.2': ['high', 'xhigh'],
 };
 
 /**

package/index.d.ts

@@ -496,7 +496,7 @@ export declare class ModelMix {
   grok420(args?: ModelAttachArgs): this;
 
   // Multi-provider
-  qwen3(args?: ModelAttachArgs): this;
+  museGlimmer30b(args?: ModelAttachArgs): this;
   qwen35397b(args?: ModelAttachArgs): this;
   qwen36plus(args?: ModelAttachArgs): this;
   qwen37plus(args?: ModelAttachArgs): this;
@@ -516,7 +516,6 @@ export declare class ModelMix {
   mimo25pro(args?: ModelAttachArgs): this;
   deepseekV4Pro(args?: ModelAttachArgs): this;
   deepseekV4Flash(args?: ModelAttachArgs): this;
-  GLM51(args?: ModelAttachArgs): this;
   GLM52(args?: ModelAttachArgs): this;
   GLM53(args?: ModelAttachArgs): this;
 

package/index.js

@@ -490,6 +490,8 @@ class ModelMix {
     }
     gptOss({ options = {}, config = {}, mix = {} } = {}) {
         mix = { ...this.mix, ...mix };
+        if (mix.nvidia) this.attach('openai/gpt-oss-120b', new MixNVIDIA({ options, config }));
+        if (mix.fireworks) this.attach('accounts/fireworks/models/gpt-oss-120b', new MixFireworks({ options, config }));
         if (mix.together) this.attach('openai/gpt-oss-120b', new MixTogether({ options, config }));
         if (mix.cerebras) this.attach('gpt-oss-120b', new MixCerebras({ options, config }));
         if (mix.groq) this.attach('openai/gpt-oss-120b', new MixGroq({ options, config }));
@@ -574,9 +576,12 @@ class ModelMix {
         return this.attach('grok-4.20-0309', new MixGrok({ options, config }));
     }
 
-    qwen3({ options = {}, config = {}, mix = { together: true, cerebras: false } } = {}) {
-        if (mix.together) this.attach('Qwen/Qwen3-235B-A22B-fp8-tput', new MixTogether({ options, config }));
-        if (mix.cerebras) this.attach('qwen-3-32b', new MixCerebras({ options, config }));
+    museGlimmer30b({ options = {}, config = {}, mix = { fireworks: true } } = {}) {
+        mix = { ...this.mix, ...mix };
+        if (mix.nvidia) this.attach('meta/muse-glimmer-30b', new MixNVIDIA({ options, config }));
+        if (mix.fireworks) this.attach('accounts/fireworks/models/muse-glimmer-30b', new MixFireworks({ options, config }));
+        if (mix.openrouter) this.attach('meta/muse-glimmer-30b', new MixOpenRouter({ options, config }));
+        if (mix.together) this.attach('meta-models/Muse-Glimmer-30B', new MixTogether({ options, config }));
         return this;
     }
 
@@ -584,9 +589,10 @@ class ModelMix {
         return this.attach('qwen/qwen3.5-397b-a17b', new MixOpenRouter({ options, config }));
     }
 
-    qwen36plus({ options = {}, config = {}, mix = { fireworks: true } } = {}) {
+    qwen36plus({ options = {}, config = {}, mix = { fireworks: false, openrouter: true } } = {}) {
         mix = { ...this.mix, ...mix };
         if (mix.fireworks) this.attach('accounts/fireworks/models/qwen3p6-plus', new MixFireworks({ options, config }));
+        if (mix.openrouter) this.attach('qwen/qwen3.6-plus', new MixOpenRouter({ options, config }));
         if (mix.together) this.attach('Qwen/Qwen3.6-Plus', new MixTogether({ options, config }));
         return this;
     }
@@ -595,6 +601,7 @@ class ModelMix {
         mix = { ...this.mix, ...mix };
         if (mix.fireworks) this.attach('accounts/fireworks/models/qwen3p7-plus', new MixFireworks({ options, config }));
         if (mix.openrouter) this.attach('qwen/qwen3.7-plus', new MixOpenRouter({ options, config }));
+        if (mix.together) this.attach('Qwen/Qwen3.7-Plus', new MixTogether({ options, config }));
         return this;
     }
 
@@ -635,13 +642,17 @@ class ModelMix {
     kimiK27Code({ options = {}, config = {}, mix = { together: true } } = {}) {
         mix = { ...this.mix, ...mix };
         if (mix.together) this.attach('moonshotai/Kimi-K2.7-Code', new MixTogether({ options, config }));
+        if (mix.fireworks) this.attach('accounts/fireworks/models/kimi-k2p7-code', new MixFireworks({ options, config }));
+        if (mix.openrouter) this.attach('moonshotai/kimi-k2.7-code', new MixOpenRouter({ options, config }));
         return this;
     }
 
     kimiK3({ options = {}, config = {}, mix = { moonshot: true, openrouter: false } } = {}) {
         mix = { ...this.mix, ...mix };
         if (mix.moonshot) this.attach('kimi-k3', new MixKimi({ options, config }));
+        if (mix.fireworks) this.attach('accounts/fireworks/models/kimi-k3', new MixFireworks({ options, config }));
         if (mix.openrouter) this.attach('moonshotai/kimi-k3', new MixOpenRouter({ options, config }));
+        if (mix.together) this.attach('moonshotai/Kimi-K3', new MixTogether({ options, config }));
         return this;
     }
 
@@ -661,14 +672,16 @@ class ModelMix {
     minimaxM27({ options = {}, config = {}, mix = { openrouter: true, minimax: true } } = {}) {
         mix = { ...this.mix, ...mix };
         if (mix.nvidia) this.attach('minimaxai/minimax-m2.7', new MixNVIDIA({ options, config }));
-        if (mix.openrouter) return this.attach('minimax/minimax-m2.7', new MixOpenRouter({ options, config }));
-        if (mix.minimax) return this.attach('MiniMax-M2.7', new MixMiniMax({ options, config }));
-        if (mix.together) return this.attach('MiniMaxAI/MiniMax-M2.7', new MixTogether({ options, config }));
+        if (mix.fireworks) this.attach('accounts/fireworks/models/minimax-m2p7', new MixFireworks({ options, config }));
+        if (mix.openrouter) this.attach('minimax/minimax-m2.7', new MixOpenRouter({ options, config }));
+        if (mix.minimax) this.attach('MiniMax-M2.7', new MixMiniMax({ options, config }));
+        if (mix.together) this.attach('MiniMaxAI/MiniMax-M2.7', new MixTogether({ options, config }));
         return this;
     }
 
     minimaxM3({ options = {}, config = {}, mix = { minimax: true, openrouter: false } } = {}) {
         mix = { ...this.mix, ...mix };
+        if (mix.fireworks) this.attach('accounts/fireworks/models/minimax-m3', new MixFireworks({ options, config }));
         if (mix.openrouter) this.attach('minimax/minimax-m3', new MixOpenRouter({ options, config }));
         if (mix.minimax) this.attach('MiniMax-M3', new MixMiniMax({ options, config }));
         if (mix.together) this.attach('MiniMaxAI/MiniMax-M3', new MixTogether({ options, config }));
@@ -707,18 +720,11 @@ class ModelMix {
         return this;
     }
 
-    GLM51({ options = {}, config = {}, mix = { fireworks: true } } = {}) {
-        mix = { ...this.mix, ...mix };
-        if (mix.nvidia) this.attach('z-ai/glm-5.1', new MixNVIDIA({ options, config }));
-        if (mix.fireworks) this.attach('accounts/fireworks/models/glm-5p1', new MixFireworks({ options, config }));
-        if (mix.openrouter) this.attach('z-ai/glm-5.1', new MixOpenRouter({ options, config }));
-        if (mix.together) this.attach('zai-org/GLM-5.1', new MixTogether({ options, config }));
-        return this;
-    }
-
     GLM52({ options = {}, config = {}, mix = { together: true } } = {}) {
         mix = { ...this.mix, ...mix };
         if (mix.together) this.attach('zai-org/GLM-5.2', new MixTogether({ options, config }));
+        if (mix.fireworks) this.attach('accounts/fireworks/models/glm-5p2', new MixFireworks({ options, config }));
+        if (mix.openrouter) this.attach('z-ai/glm-5.2', new MixOpenRouter({ options, config }));
         return this;
     }
 

package/lib/model-chain.js

@@ -10,11 +10,12 @@ const CHAIN_MODEL_SHORTCUTS = new Set([
     'gemini31pro', 'gemini37flash', 'gemini36flash', 'gemini35flash',
     'gemini35flashLite', 'gemini31flashLite', 'sonarPro', 'sonar',
     'grok46', 'grok45', 'grok43', 'grok420multiAgent', 'grok420',
-    'qwen3', 'qwen35397b', 'qwen36plus', 'qwen37plus', 'qwen38max', 'qwen3827b',
+    'museGlimmer30b',
+    'qwen35397b', 'qwen36plus', 'qwen37plus', 'qwen38max', 'qwen3827b',
     'hermes470b', 'hermes4405b', 'hermes3',
     'kimiK26', 'kimiK27Code', 'kimiK3', 'kimiK25',
     'minimaxM27', 'minimaxM3', 'mimo25', 'mimo25pro',
-    'deepseekV4Pro', 'deepseekV4Flash', 'GLM51', 'GLM52', 'GLM53'
+    'deepseekV4Pro', 'deepseekV4Flash', 'GLM52', 'GLM53'
 ]);
 
 function parseChainModels(modelSpecs) {

package/lib/token-usage.js

@@ -42,9 +42,10 @@ const MODEL_PRICING = {
     'gpt-4.1': { input: 2.00, cachedInput: 0.50, output: 8.00 },
     'gpt-4.1-mini': { input: 0.40, cachedInput: 0.10, output: 1.60 },
     'gpt-4.1-nano': { input: 0.10, cachedInput: 0.025, output: 0.40 },
-    // gptOss (Together/Groq/Cerebras/OpenRouter)
+    // gptOss (NVIDIA/Fireworks/Together/Groq/Cerebras/OpenRouter)
     'openai/gpt-oss-120b': { input: 0.15, output: 0.60 },
     'gpt-oss-120b': { input: 0.15, output: 0.60 },
+    'accounts/fireworks/models/gpt-oss-120b': { input: 0.15, cachedInput: 0.014, output: 0.60 },
     // Anthropic
     'claude-fable-5': { input: 10.00, cachedInput: 1.00, cacheWrite: 12.50, cacheWrite1h: 20.00, output: 50.00 },
     'claude-opus-5': { input: 5.00, cachedInput: 0.50, cacheWrite: 6.25, cacheWrite1h: 10.00, output: 25.00 },
@@ -74,6 +75,10 @@ const MODEL_PRICING = {
     'grok-4.20-0309': { input: 1.25, output: 2.50 },
     'grok-4.20-0309-reasoning': { input: 1.25, output: 2.50 },
     'grok-4.20-0309-non-reasoning': { input: 1.25, output: 2.50 },
+    // Muse Glimmer 30B (NVIDIA/OpenRouter, Fireworks, Together)
+    'meta/muse-glimmer-30b': { input: 0.35, cachedInput: 0.04, output: 1.50 },
+    'accounts/fireworks/models/muse-glimmer-30b': { input: 0.35, cachedInput: 0.04, output: 1.50 },
+    'meta-models/Muse-Glimmer-30B': { input: 0.35, cachedInput: 0.04, output: 1.50 },
     // Fireworks
     'accounts/fireworks/models/deepseek-v4-flash': { input: 0.14, output: 0.28 },
     'accounts/fireworks/models/deepseek-v4-pro': { input: 1.74, output: 3.48 },
@@ -82,15 +87,29 @@ const MODEL_PRICING = {
     'deepseek-ai/DeepSeek-V4-Pro': { input: 2.10, output: 4.40 },
     'deepseek/deepseek-v4-flash': { input: 0.09, output: 0.18 },
     'accounts/fireworks/models/glm-4p7': { input: 0.55, output: 2.19 },
-    'accounts/fireworks/models/glm-5p1': { input: 1.05, output: 3.50 },
-    'zai-org/GLM-5.2': { input: 1.40, output: 4.40 },
+    'zai-org/GLM-5.2': { input: 1.40, cachedInput: 0.26, output: 4.40 },
+    'accounts/fireworks/models/glm-5p2': { input: 1.40, cachedInput: 0.14, output: 4.40 },
+    'z-ai/glm-5.2': { input: 0.966, cachedInput: 0.1932, output: 3.036 },
     'z-ai/glm-5.3': { input: 1.40, cachedInput: 0.26, output: 4.40 },
     'accounts/fireworks/models/kimi-k2p5': { input: 0.50, output: 2.80 },
     'qwen/qwen3.5-397b-a17b': { input: 0.385, output: 2.45 },
     'accounts/fireworks/models/qwen3p6-plus': { input: 0.50, output: 3.00 },
     'Qwen/Qwen3.6-Plus': { input: 0.50, output: 3.00 },
+    'qwen/qwen3.6-plus': {
+        input: 0.325,
+        cacheWrite: 0.40625,
+        output: 1.95,
+        longContext: { inputThreshold: 256_000, inputMultiplier: 4, outputMultiplier: 2, inclusive: true }
+    },
     'accounts/fireworks/models/qwen3p7-plus': { input: 0.40, output: 1.60 },
-    'qwen/qwen3.7-plus': { input: 0.32, output: 1.28 },
+    'qwen/qwen3.7-plus': {
+        input: 0.32,
+        cachedInput: 0.064,
+        cacheWrite: 0.40,
+        output: 1.28,
+        longContext: { inputThreshold: 256_000, inputMultiplier: 3, outputMultiplier: 3, inclusive: true }
+    },
+    'Qwen/Qwen3.7-Plus': { input: 0.32, output: 1.28 },
     'accounts/fireworks/models/qwen3p8-2p4t-a95b': { input: 2.00, cachedInput: 0.25, output: 6.00 },
     'qwen/qwen3.8-max': { input: 2.00, output: 6.00 },
     'qwen/qwen3.8-27b': { input: 0.45, cachedInput: 0.05, output: 3.20 },
@@ -100,7 +119,11 @@ const MODEL_PRICING = {
     'MiniMax-M3': { input: 0.30, output: 1.20 },
     'minimax/minimax-m2.7': { input: 0.30, output: 1.20 },
     'minimax/minimax-m3': { input: 0.30, output: 1.20 },
+    'minimaxai/minimax-m2.7': { input: 0.30, output: 1.20 },
+    'MiniMaxAI/MiniMax-M2.7': { input: 0.30, output: 1.20 },
     'MiniMaxAI/MiniMax-M3': { input: 0.30, output: 1.20 },
+    'accounts/fireworks/models/minimax-m2p7': { input: 0.30, cachedInput: 0.059, output: 1.20 },
+    'accounts/fireworks/models/minimax-m3': { input: 0.30, cachedInput: 0.059, output: 1.20 },
     // Perplexity
     'sonar': { input: 1.00, output: 1.00 },
     'sonar-pro': { input: 3.00, output: 15.00 },
@@ -110,15 +133,18 @@ const MODEL_PRICING = {
     // Hermes 3 (Lambda/OpenRouter)
     'Hermes-3-Llama-3.1-405B-FP8': { input: 0.80, output: 0.80 },
     'nousresearch/hermes-3-llama-3.1-405b:free': { input: 0, output: 0 },
-    // Qwen3 (Together/Cerebras)
-    'Qwen/Qwen3-235B-A22B-fp8-tput': { input: 0.20, output: 0.60 },
-    'qwen-3-32b': { input: 0.20, output: 0.60 },
     // Kimi K2.5 (Together/Fireworks/OpenRouter)
     'moonshotai/Kimi-K2.5': { input: 0.50, output: 2.80 },
     'moonshotai/kimi-k2.5': { input: 0.50, output: 2.80 },
     // Kimi K3
-    'kimi-k3': { input: 3.00, output: 15.00 },
-    'moonshotai/kimi-k3': { input: 3.00, output: 15.00 },
+    'kimi-k3': { input: 3.00, cachedInput: 0.30, output: 15.00 },
+    'moonshotai/kimi-k3': { input: 3.00, cachedInput: 0.30, output: 15.00 },
+    'accounts/fireworks/models/kimi-k3': { input: 3.00, cachedInput: 0.30, output: 15.00 },
+    'moonshotai/Kimi-K3': { input: 3.00, cachedInput: 0.30, output: 15.00 },
+    // Kimi K2.7 Code
+    'moonshotai/Kimi-K2.7-Code': { input: 0.95, cachedInput: 0.19, output: 4.00 },
+    'accounts/fireworks/models/kimi-k2p7-code': { input: 0.95, cachedInput: 0.19, output: 4.00 },
+    'moonshotai/kimi-k2.7-code': { input: 0.71, cachedInput: 0.15, output: 3.50 },
     // GLM 4.7 (OpenRouter/Cerebras)
     'z-ai/glm-4.7': { input: 0.55, output: 2.19 },
     'zai-glm-4.7': { input: 0.55, output: 2.19 },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "modelmix",
-  "version": "5.1.4",
+  "version": "5.1.6",
   "description": "🧬 Reliable interface with automatic fallback for AI LLMs.",
   "main": "index.js",
   "types": "index.d.ts",

package/skills/modelmix/SKILL.md

@@ -168,22 +168,25 @@ Use `.effort(n)` (or `config.effort`) to enable Anthropic thinking — e.g. `.ef
 `sonar()` `sonarPro()`
 
 ### Together
-`qwen36plus()` `GLM52()` `kimiK25()` `kimiK26()` `gptOss()`
+`museGlimmer30b()` `qwen36plus()` `qwen37plus()` `GLM52()` `kimiK25()` `kimiK26()` `kimiK27Code()` `kimiK3()` `minimaxM27()` `minimaxM3()` `gptOss()`
+
+### NVIDIA
+`museGlimmer30b()` `gptOss()` `minimaxM27()`
 
 ### Moonshot
-`kimiK3()` — requires `MOONSHOT_API_KEY`; use `{ mix: { moonshot: false, openrouter: true } }` for OpenRouter.
+`kimiK3()` — requires `MOONSHOT_API_KEY`; Fireworks, OpenRouter, and Together are available through `mix`.
 
 ### MiniMax
 `minimaxM27()` `minimaxM3()`
 
 ### Fireworks
-`qwen36plus()` `qwen37plus()` `qwen38max()` `deepseekV4Flash()` `deepseekV4Pro()` `kimiK26()`
+`museGlimmer30b()` `gptOss()` `qwen36plus()` (private/on-demand only) `qwen37plus()` `qwen38max()` `deepseekV4Flash()` `deepseekV4Pro()` `kimiK26()` `kimiK27Code()` `kimiK3()` `minimaxM27()` `minimaxM3()` `GLM52()`
 
 ### Cerebras
 `GLM46()`
 
 ### OpenRouter
-`qwen35397b()` `qwen3827b()` `hermes470b()` `hermes4405b()` `qwen38max()` `GLM45()` `GLM53()`
+`museGlimmer30b()` `gptOss()` `qwen35397b()` `qwen36plus()` `qwen37plus()` `qwen3827b()` `hermes470b()` `hermes4405b()` `qwen38max()` `kimiK27Code()` `kimiK3()` `minimaxM27()` `minimaxM3()` `GLM45()` `GLM52()` `GLM53()`
 
 ### Multi-provider (auto-fallback across free/paid tiers)
 `hermes3()` `kimiK25()`

package/test/effort.test.js

@@ -96,6 +96,42 @@ describe('Unified effort scale', () => {
             expect(mapEffort('openai', 100, key)).to.deep.equal({ reasoning_effort: 'high' });
         });
 
+        it('maps the new GPT OSS providers to low, medium, and high reasoning', () => {
+            expect(mapEffort('openai', 0, 'accounts/fireworks/models/gpt-oss-120b')).to.deep.equal({
+                reasoning_effort: 'low'
+            });
+            expect(mapEffort('openai', 50, 'openai/gpt-oss-120b')).to.deep.equal({
+                reasoning_effort: 'medium'
+            });
+            expect(mapEffort('openai', 100, 'accounts/fireworks/models/gpt-oss-120b')).to.deep.equal({
+                reasoning_effort: 'high'
+            });
+        });
+
+        it('maps Kimi K3 and GLM 5.2 to their provider-supported levels', () => {
+            expect(mapEffort('openai', 100, 'moonshotai/Kimi-K3')).to.deep.equal({
+                reasoning_effort: 'max'
+            });
+            expect(mapEffort('openai', 0, 'accounts/fireworks/models/glm-5p2')).to.deep.equal({
+                reasoning_effort: 'high'
+            });
+            expect(mapEffort('openai', 100, 'z-ai/glm-5.2')).to.deep.equal({
+                reasoning_effort: 'xhigh'
+            });
+        });
+
+        it('maps Muse Glimmer 30B to provider-supported reasoning levels', () => {
+            expect(mapEffort('openai', 0, 'accounts/fireworks/models/muse-glimmer-30b')).to.deep.equal({
+                reasoning_effort: 'low'
+            });
+            expect(mapEffort('openai', 100, 'accounts/fireworks/models/muse-glimmer-30b')).to.deep.equal({
+                reasoning_effort: 'xhigh'
+            });
+            expect(mapEffort('openai', 100, 'meta/muse-glimmer-30b')).to.deep.equal({
+                reasoning_effort: 'high'
+            });
+        });
+
         it('maps Qwen 3.8 27B to its supported reasoning levels', () => {
             const key = 'qwen/qwen3.8-27b';
             expect(mapEffort('openai', 0, key)).to.deep.equal({ reasoning_effort: 'low' });

package/test/fallback.test.js

@@ -64,7 +64,7 @@ describe('Provider Fallback Chain Tests', () => {
             for (const shortcut of [
                 'gpt41', 'gpt41mini', 'gpt41nano',
                 'gemini25flash', 'gemini25pro', 'gemini3pro', 'gemini3flash',
-                'minimaxM25'
+                'minimaxM25', 'qwen3', 'GLM51'
             ]) {
                 expect(model[shortcut]).to.equal(undefined);
                 expect(() => model.chain(shortcut))

package/test/glm.test.js

@@ -1,13 +1,37 @@
 const { expect } = require('chai');
-const { ModelMix, MixOpenRouter } = require('../index.js');
+const { ModelMix, MixFireworks, MixOpenRouter, MixTogether } = require('../index.js');
 
 describe('GLM Model Registration Tests', () => {
-    it('should register Together GLM 5.2 by default', () => {
+    it('should register Together GLM 5.2 before the OpenRouter fallback by default', () => {
         const model = ModelMix.new();
         model.GLM52();
 
-        expect(model.models).to.have.length(1);
-        expect(model.models[0].key).to.equal('zai-org/GLM-5.2');
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'zai-org/GLM-5.2',
+            'z-ai/glm-5.2'
+        ]);
+        expect(model.models[0].provider).to.be.instanceOf(MixTogether);
+        expect(model.models[1].provider).to.be.instanceOf(MixOpenRouter);
+    });
+
+    it('should register every requested GLM 5.2 provider', () => {
+        const model = ModelMix.new().GLM52({
+            mix: { together: true, fireworks: true, openrouter: true }
+        });
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'zai-org/GLM-5.2',
+            'accounts/fireworks/models/glm-5p2',
+            'z-ai/glm-5.2'
+        ]);
+        expect(model.models[0].provider).to.be.instanceOf(MixTogether);
+        expect(model.models[1].provider).to.be.instanceOf(MixFireworks);
+        expect(model.models[2].provider).to.be.instanceOf(MixOpenRouter);
+        expect(ModelMix.calculateCost('accounts/fireworks/models/glm-5p2', {
+            input: 1_000_000,
+            cached: 500_000,
+            output: 1_000_000
+        })).to.equal(5.17);
     });
 
     it('should register GLM 5.3 through OpenRouter', () => {

package/test/kimi.test.js

@@ -1,15 +1,32 @@
 const { expect } = require('chai');
 const nock = require('nock');
-const { ModelMix, MixKimi, MixOpenRouter, MixTogether } = require('../index.js');
+const { ModelMix, MixFireworks, MixKimi, MixOpenRouter, MixTogether } = require('../index.js');
 
 describe('Kimi Model Registration Tests', () => {
-    it('should register Together Kimi K2.7 Code by default', () => {
+    it('should register Together Kimi K2.7 Code before the OpenRouter fallback by default', () => {
         const model = ModelMix.new();
         model.kimiK27Code();
 
-        expect(model.models).to.have.length(1);
-        expect(model.models[0].key).to.equal('moonshotai/Kimi-K2.7-Code');
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'moonshotai/Kimi-K2.7-Code',
+            'moonshotai/kimi-k2.7-code'
+        ]);
         expect(model.models[0].provider).to.be.instanceOf(MixTogether);
+        expect(model.models[1].provider).to.be.instanceOf(MixOpenRouter);
+    });
+
+    it('should register every requested Kimi K2.7 Code provider', () => {
+        const model = ModelMix.new().kimiK27Code({
+            mix: { together: true, fireworks: true, openrouter: true }
+        });
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'moonshotai/Kimi-K2.7-Code',
+            'accounts/fireworks/models/kimi-k2p7-code',
+            'moonshotai/kimi-k2.7-code'
+        ]);
+        expect(model.models[1].provider).to.be.instanceOf(MixFireworks);
+        expect(model.models[2].provider).to.be.instanceOf(MixOpenRouter);
     });
 
     it('should register Kimi K3 with the native Moonshot provider by default', () => {
@@ -45,6 +62,19 @@ describe('Kimi Model Registration Tests', () => {
         }
     });
 
+    it('should register Fireworks and Together Kimi K3 when requested', () => {
+        const model = ModelMix.new().kimiK3({
+            mix: { moonshot: false, fireworks: true, openrouter: false, together: true }
+        });
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'accounts/fireworks/models/kimi-k3',
+            'moonshotai/Kimi-K3'
+        ]);
+        expect(model.models[0].provider).to.be.instanceOf(MixFireworks);
+        expect(model.models[1].provider).to.be.instanceOf(MixTogether);
+    });
+
     it('should adapt Kimi K3 requests to its fixed sampling API', async () => {
         const originalMoonshotApiKey = process.env.MOONSHOT_API_KEY;
         process.env.MOONSHOT_API_KEY = 'test-moonshot-key';

package/test/muse.test.js

@@ -0,0 +1,61 @@
+const { expect } = require('chai');
+const {
+    ModelMix,
+    MixFireworks,
+    MixNVIDIA,
+    MixOpenRouter,
+    MixTogether
+} = require('../index.js');
+
+describe('Muse Glimmer Model Registration Tests', () => {
+    it('registers Fireworks before the OpenRouter fallback by default', () => {
+        const model = ModelMix.new().museGlimmer30b();
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'accounts/fireworks/models/muse-glimmer-30b',
+            'meta/muse-glimmer-30b'
+        ]);
+        expect(model.models[0].provider).to.be.instanceOf(MixFireworks);
+        expect(model.models[1].provider).to.be.instanceOf(MixOpenRouter);
+    });
+
+    it('registers every supported provider in fallback order', () => {
+        const model = ModelMix.new().museGlimmer30b({
+            mix: { nvidia: true, fireworks: true, openrouter: true, together: true }
+        });
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'meta/muse-glimmer-30b',
+            'accounts/fireworks/models/muse-glimmer-30b',
+            'meta/muse-glimmer-30b',
+            'meta-models/Muse-Glimmer-30B'
+        ]);
+        expect(model.models.map(({ provider }) => provider.constructor)).to.deep.equal([
+            MixNVIDIA,
+            MixFireworks,
+            MixOpenRouter,
+            MixTogether
+        ]);
+    });
+
+    it('calculates cached input pricing for every provider model ID', () => {
+        const usage = { input: 1_000_000, cached: 250_000, output: 1_000_000 };
+
+        for (const key of [
+            'meta/muse-glimmer-30b',
+            'accounts/fireworks/models/muse-glimmer-30b',
+            'meta-models/Muse-Glimmer-30B'
+        ]) {
+            expect(ModelMix.calculateCost(key, usage)).to.equal(1.7725);
+        }
+    });
+
+    it('supports Muse Glimmer 30B in chain()', () => {
+        const model = ModelMix.new().chain('museGlimmer30b');
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'accounts/fireworks/models/muse-glimmer-30b',
+            'meta/muse-glimmer-30b'
+        ]);
+    });
+});

package/test/provider-expansion.test.js

@@ -0,0 +1,83 @@
+const { expect } = require('chai');
+const {
+    ModelMix,
+    MixFireworks,
+    MixMiniMax,
+    MixNVIDIA,
+    MixOpenRouter,
+    MixTogether
+} = require('../index.js');
+
+describe('Provider expansion regressions', () => {
+    it('should retain every enabled GPT OSS provider, including shared model keys', () => {
+        const model = ModelMix.new().gptOss({
+            mix: {
+                nvidia: true,
+                fireworks: true,
+                together: true,
+                cerebras: true,
+                groq: true,
+                openrouter: true
+            }
+        });
+
+        expect(model.models.map(({ key }) => key)).to.deep.equal([
+            'openai/gpt-oss-120b',
+            'accounts/fireworks/models/gpt-oss-120b',
+            'openai/gpt-oss-120b',
+            'gpt-oss-120b',
+            'openai/gpt-oss-120b',
+            'openai/gpt-oss-120b'
+        ]);
+        expect(model.models[0].provider).to.be.instanceOf(MixNVIDIA);
+        expect(model.models[1].provider).to.be.instanceOf(MixFireworks);
+        expect(ModelMix.calculateCost('accounts/fireworks/models/gpt-oss-120b', {
+            input: 1_000_000,
+            cached: 500_000,
+            output: 1_000_000
+        })).to.equal(0.682);
+    });
+
+    it('should retain every enabled MiniMax M2.7 provider in fallback order', () => {
+        const originalMiniMaxApiKey = process.env.MINIMAX_API_KEY;
+        process.env.MINIMAX_API_KEY = 'test-minimax-key';
+
+        try {
+            const model = ModelMix.new().minimaxM27({
+                mix: {
+                    nvidia: true,
+                    fireworks: true,
+                    openrouter: true,
+                    minimax: true,
+                    together: true
+                }
+            });
+
+            expect(model.models.map(({ key }) => key)).to.deep.equal([
+                'minimaxai/minimax-m2.7',
+                'accounts/fireworks/models/minimax-m2p7',
+                'minimax/minimax-m2.7',
+                'MiniMax-M2.7',
+                'MiniMaxAI/MiniMax-M2.7'
+            ]);
+            expect(model.models[0].provider).to.be.instanceOf(MixNVIDIA);
+            expect(model.models[1].provider).to.be.instanceOf(MixFireworks);
+            expect(model.models[2].provider).to.be.instanceOf(MixOpenRouter);
+            expect(model.models[3].provider).to.be.instanceOf(MixMiniMax);
+            expect(model.models[4].provider).to.be.instanceOf(MixTogether);
+        } finally {
+            if (originalMiniMaxApiKey === undefined) delete process.env.MINIMAX_API_KEY;
+            else process.env.MINIMAX_API_KEY = originalMiniMaxApiKey;
+        }
+    });
+
+    it('should register Fireworks for MiniMax M3 without changing its default', () => {
+        const model = ModelMix.new().minimaxM3({
+            mix: { fireworks: true, openrouter: false, minimax: false, together: false }
+        });
+
+        expect(model.models).to.have.length(1);
+        expect(model.models[0].key).to.equal('accounts/fireworks/models/minimax-m3');
+        expect(model.models[0].provider).to.be.instanceOf(MixFireworks);
+    });
+});

package/test/qwen.test.js

@@ -1,13 +1,22 @@
 const { expect } = require('chai');
-const { ModelMix, MixOpenRouter } = require('../index.js');
+const { ModelMix, MixFireworks, MixOpenRouter, MixTogether } = require('../index.js');
 
 describe('Qwen Model Registration Tests', () => {
-    it('should register Fireworks Qwen 3.6 Plus by default', () => {
+    it('should register OpenRouter Qwen 3.6 Plus by default', () => {
+        const model = ModelMix.new().qwen36plus();
+
+        expect(model.models).to.have.length(1);
+        expect(model.models[0].key).to.equal('qwen/qwen3.6-plus');
+        expect(model.models[0].provider).to.be.instanceOf(MixOpenRouter);
+    });
+
+    it('should retain Fireworks Qwen 3.6 Plus as an explicit deployment option', () => {
         const model = ModelMix.new();
-        model.qwen36plus({ mix: { fireworks: true, together: false } });
+        model.qwen36plus({ mix: { fireworks: true, openrouter: false, together: false } });
 
         expect(model.models).to.have.length(1);
         expect(model.models[0].key).to.equal('accounts/fireworks/models/qwen3p6-plus');
+        expect(model.models[0].provider).to.be.instanceOf(MixFireworks);
     });
 
     it('should register Together Qwen 3.6 Plus when together mix is enabled', () => {
@@ -38,6 +47,26 @@ describe('Qwen Model Registration Tests', () => {
         expect(model.models[0].key).to.equal('qwen/qwen3.7-plus');
     });
 
+    it('should register Together Qwen 3.7 Plus when together mix is enabled', () => {
+        const model = ModelMix.new();
+        model.qwen37plus({ mix: { fireworks: false, openrouter: false, together: true } });
+
+        expect(model.models).to.have.length(1);
+        expect(model.models[0].key).to.equal('Qwen/Qwen3.7-Plus');
+        expect(model.models[0].provider).to.be.instanceOf(MixTogether);
+    });
+
+    it('should apply OpenRouter long-context pricing to Qwen 3.6 and 3.7 Plus', () => {
+        expect(ModelMix.calculateCost('qwen/qwen3.6-plus', {
+            input: 256_000,
+            output: 1_000_000
+        })).to.equal(4.2328);
+        expect(ModelMix.calculateCost('qwen/qwen3.7-plus', {
+            input: 256_000,
+            output: 1_000_000
+        })).to.equal(4.08576);
+    });
+
     it('should register Fireworks Qwen 3.8 Max before the OpenRouter fallback by default', () => {
         const model = ModelMix.new();
         model.qwen38max();