vtcode_config/models/
model_id.rs

1use serde::{Deserialize, Serialize};
2
3mod as_str;
4mod capabilities;
5mod collection;
6mod defaults;
7mod description;
8mod display;
9mod format;
10mod openrouter;
11mod parse;
12mod provider;
13mod table;
14
15pub use capabilities::{
16    ModelCatalogEntry, ModelPricing, catalog_provider_keys, model_catalog_entry, supported_models_for_provider,
17};
18
19/// Centralized enum for all supported model identifiers
20#[cfg_attr(feature = "schema", derive(schemars::JsonSchema))]
21#[derive(Clone, Debug, Default, PartialEq, Eq, Hash, Serialize, Deserialize)]
22pub enum ModelId {
23    // Gemini models
24    /// Gemini 3.6 Flash - Latest flash model with improved capabilities
25    /// Gemini 3.7 Flash - Flash model with 1M context and tunable thinking levels
26    /// Gemini 3.8 Flash - Most intelligent Flash for long-horizon SWE, agents, and enterprise workflows (1M context, 64k output, low/medium/high thinking)
27    Gemini38Flash,
28    /// Gemini 4 Argon - Flagship reasoning model for the hardest coding, agentic, math, and science problems (unified 1M context window, 64K sync output, long decode, low/medium/high thinking)
29    Gemini4Argon,
30
31    // OpenAI models
32    /// GPT-6 Astra - Most capable model for hardest end-to-end work with complex reasoning, coding, computer use, and research
33    GPT6Astra,
34    /// GPT-6 Sol - Cost-efficient high-end model in the GPT-6 series for demanding professional work
35    GPT6Sol,
36    /// GPT-6.1 Sol - Near-Astra performance for complex coding, computer use, and professional work at a lower cost
37    GPT61Sol,
38    /// GPT-6.1 Terra - Fast, efficient model for visual reasoning, precise computer use, and clear, natural writing
39    GPT61Terra,
40    /// GPT-6 Luna - Fast cost-efficient model in the GPT-6 series for high-volume latency-sensitive workloads
41    GPT6Luna,
42    /// GPT-5.6 Sol - Frontier model for complex professional work in the GPT-5.6 family
43    GPT56Sol,
44    /// GPT-5.6 Terra - GPT-5.6 model that balances intelligence and cost
45    GPT56Terra,
46    /// GPT-5.6 Luna - GPT-5.6 model optimized for cost-sensitive workloads
47    GPT56Luna,
48    /// GPT-OSS 20B - OpenAI's open-source 20B parameter model using harmony
49    OpenAIGptOss20b,
50    /// GPT-OSS 120B - OpenAI's open-source 120B parameter model using harmony
51    OpenAIGptOss120b,
52
53    // Anthropic models
54    /// Claude Sonnet 5 - The best combination of speed and intelligence with adaptive thinking on by default
55    #[default]
56    ClaudeSonnet5,
57    /// Claude Sonnet 5.5 - Latest Sonnet: best speed/intelligence balance, 1M context, 128K output, `between_tools` as the lowest thinking setting, default effort high
58    ClaudeSonnet55,
59    /// Claude Haiku 5.5 - High-volume latency-sensitive work: classification, routing, extraction, subagents; adaptive thinking, 1M context, 128K output, default effort medium
60    ClaudeHaiku55,
61    /// Claude Fable 5 - Anthropic's most capable widely released model for demanding reasoning and long-horizon agentic work
62    ClaudeFable5,
63    /// Claude Fable 5.1 - successor to Fable 5 for demanding reasoning and long-horizon agentic work, 1M context, adaptive thinking always on, cache reads at 1/4 cost
64    ClaudeFable51,
65    /// Claude Fable 5.2 - latest Fable for demanding reasoning and long-horizon agentic work, 1M context, 128K output, adaptive thinking always on, default effort high
66    ClaudeFable52,
67    /// Claude Opus 5 - Anthropic's newest Opus-tier model with 1M context, thinking on by default
68    ClaudeOpus5,
69    /// Claude Opus 5.5 - Opus-tier successor for long-running agentic coding, adaptive thinking always on, 1M context, 128K output, default effort medium
70    ClaudeOpus55,
71    /// GitHub Copilot auto model selection
72    CopilotAuto,
73    /// GitHub Copilot GPT-5.2 Codex
74    CopilotGPT52Codex,
75    /// GitHub Copilot GPT-5.1 Codex Max
76    CopilotGPT51CodexMax,
77    /// GitHub Copilot GPT-5.4
78    CopilotGPT54,
79    /// GitHub Copilot GPT-5.4 Mini
80    CopilotGPT54Mini,
81
82    // DeepSeek models
83    /// DeepSeek V4.1 Flash - Latest flash model with improved reasoning and efficiency
84    DeepSeekFlash,
85
86    // Official Meta AI models
87    /// Meta Muse Spark 1.3 - Official Meta AI flagship Standard-tier reasoning model, tuned for agentic workflows
88    MetaMuseSpark13,
89    /// Meta Muse Spark 1.3 Contributor tier - opt-in variant with Meta's discounted data-contribution terms
90    MetaMuseSpark13Contributor,
91    /// Meta Muse Wonder 1.0 - Frontier coding, long-running agents, and advanced computer use
92    MetaMuseWonder10,
93    /// Meta Muse Wonder Full 1.0 - Most advanced Meta model for critical workflows (Muse Code subscription)
94    MetaMuseWonderFull10,
95    /// Meta Muse Ember 1.0 - Efficient execution model for high-volume sub-agent tasks, tool calls, and routing
96    MetaMuseEmber10,
97
98    // NVIDIA NIM models
99    /// NVIDIA Nemotron 3 Ultra - NVIDIA's flagship agentic reasoning model via NIM
100    NvidiaNemotron3Ultra550bA55b,
101    /// NVIDIA Nemotron 3 Super - Efficient long-context agentic reasoning model via NIM
102    NvidiaNemotron3Super120bA12b,
103    /// NVIDIA Nemotron 3 Nano - Efficient reasoning and tool-use model via NIM
104    NvidiaNemotron3Nano30bA3b,
105
106    // Merge Gateway routes
107    /// Merge Gateway's default route selected by Merge
108    MergeGatewayDefaultRouting,
109    /// Anthropic Claude Opus 5 through Merge Gateway
110    MergeGatewayAnthropicClaudeOpus5,
111    /// Anthropic Claude Opus 5.5 through Merge Gateway
112    MergeGatewayAnthropicClaudeOpus55,
113    /// Anthropic Claude Sonnet 5 through Merge Gateway
114    MergeGatewayAnthropicClaudeSonnet5,
115    /// Anthropic Claude Sonnet 5.5 through Merge Gateway
116    MergeGatewayAnthropicClaudeSonnet55,
117    /// Google Gemini 3.6 Flash through Merge Gateway
118    /// Google Gemini 3.7 Flash through Merge Gateway
119    /// DeepSeek V4.1 Flash through Merge Gateway
120    MergeGatewayDeepseekFlash,
121    /// xAI Grok 4.6 through Merge Gateway
122    MergeGatewayXaiGrok46,
123    /// xAI Grok 4.7 through Merge Gateway
124    MergeGatewayXaiGrok47,
125    /// MiniMax H3 through Merge Gateway
126    MergeGatewayMinimaxH3,
127    /// MiniMax M3.1 Flash through Merge Gateway
128    MergeGatewayMinimaxM31Flash,
129    /// Moonshot Kimi K3 through Merge Gateway
130    MergeGatewayMoonshotKimiK3,
131    /// Moonshot Kimi K3.1 Pro through Merge Gateway
132    MergeGatewayMoonshotKimiK31Pro,
133    /// Thinking Machines Inkling through Merge Gateway
134    MergeGatewayThinkingMachinesInkling,
135    /// Z.AI GLM-5.3 Flash through Merge Gateway
136    MergeGatewayZaiGlm53Flash,
137    /// Z.AI GLM-5.3 FlashX through Merge Gateway
138    MergeGatewayZaiGlm53Flashx,
139    /// OpenAI GPT-5.6 Luna through Merge Gateway
140    MergeGatewayOpenAIGpt56Luna,
141    /// OpenAI GPT-5.6 Sol through Merge Gateway
142    MergeGatewayOpenAIGpt56Sol,
143    /// OpenAI GPT-5.6 Terra through Merge Gateway
144    MergeGatewayOpenAIGpt56Terra,
145    /// Google Gemini 3.8 Flash through Merge Gateway
146    MergeGatewayGoogleGemini38Flash,
147    /// Google Gemini 4 Argon through Merge Gateway
148    MergeGatewayGoogleGemini4Argon,
149    /// Anthropic Claude Haiku 4.5 through Merge Gateway
150    MergeGatewayAnthropicClaudeHaiku4520251001,
151    /// Anthropic Claude Haiku 5.5 through Merge Gateway
152    MergeGatewayAnthropicClaudeHaiku55,
153    /// Anthropic Claude Fable 5.1 through Merge Gateway
154    MergeGatewayAnthropicClaudeFable51,
155    /// Anthropic Claude Fable 5.2 through Merge Gateway
156    MergeGatewayAnthropicClaudeFable52,
157    /// OpenAI GPT-6 Astra through Merge Gateway
158    MergeGatewayOpenAIGpt6Astra,
159    /// OpenAI GPT-6 Sol through Merge Gateway
160    MergeGatewayOpenAIGpt6Sol,
161    /// OpenAI GPT-6.1 Sol through Merge Gateway
162    MergeGatewayOpenAIGpt61Sol,
163    /// OpenAI GPT-6.1 Terra through Merge Gateway
164    MergeGatewayOpenAIGpt61Terra,
165    /// OpenAI GPT-6 Luna through Merge Gateway
166    MergeGatewayOpenAIGpt6Luna,
167    /// Xiaomi MiMo V2.6 Pro through Merge Gateway
168    MergeGatewayXiaomimimoMimoV26Pro,
169    /// Xiaomi MiMo V2.6 Flash through Merge Gateway
170    MergeGatewayXiaomimimoMimoV26Flash,
171    /// Mistral Large 4 through Merge Gateway
172    MergeGatewayMistralLarge4,
173    /// Meta Muse Spark Open (816B/42B) through Merge Gateway
174    MergeGatewayMetaMuseSparkOpen,
175
176    // Mistral AI models
177    /// Mistral Large 3 - State-of-the-art open-weight general-purpose multimodal model
178    MistralLarge3,
179    /// Mistral Large 4 - Open-weight MoE flagship (49B active / 1.05T total) with 1M context (Public Preview)
180    MistralLarge4,
181    // Hugging Face models
182    /// OpenAI GPT-OSS 20B via Hugging Face router
183    HuggingFaceOpenAIGptOss20b,
184    /// OpenAI GPT-OSS 120B via Hugging Face router
185    HuggingFaceOpenAIGptOss120b,
186    /// Z.AI GLM-5.2 via Novita inference provider on Hugging Face router
187    /// Z.AI GLM-5.3 Flash via Together inference provider on Hugging Face router
188    HuggingFaceGlm53FlashTogether,
189    /// Z.AI GLM-5.3 via Together inference provider on Hugging Face router
190    HuggingFaceGlm53Together,
191    /// Kimi K3 via Together on Hugging Face router
192    HuggingFaceKimiK3Together,
193    /// MiniMax M3 via Novita on Hugging Face router
194    HuggingFaceMinimaxM3Novita,
195    /// Meta Muse Spark Open 816B-A42B via Baseten on Hugging Face router
196    HuggingFaceMuseSparkOpenBaseten,
197
198    // StepFun models
199    /// Step 3.7 Flash - StepFun's flagship multimodal reasoning model with tool calling
200    StepFun37Flash,
201    /// Step 5 Preview - StepFun's frontier model for production-scale Agent applications with 1M context
202    StepFun5Preview,
203
204    // Evolink gateway models (namespaced as `evolink/<model>`)
205    /// GPT-5.2 served through the Evolink gateway
206    /// GPT-5.5 served through the Evolink gateway
207    /// Gemini 3.1 Pro served through the Evolink gateway (OpenAI SDK format)
208    EvolinkGemini31Pro,
209    /// Gemini 3.5 Flash served through the Evolink gateway (OpenAI SDK format)
210    /// MiniMax-M3 served through the Evolink gateway (OpenAI Chat Completions format)
211    EvolinkMinimaxM3,
212    /// Claude Haiku 4.5 served through the Evolink gateway (Anthropic Messages API)
213    EvolinkClaudeHaiku45,
214
215    /// GLM-5.3 - Z.ai flagship coding model with frontier long-horizon agentic performance
216    ZaiGlm53,
217    /// GLM-5.3 Flash - Z.ai efficient multimodal model with hybrid sparse+linear attention, 320B total / 18B active, 1M context, native vision
218    ZaiGlm53Flash,
219    /// GLM-5.3 FlashX - Z.ai high-speed Flash variant with faster inference (up to 200 tok/s), 320B total / 18B active, 1M context, native vision
220    ZaiGlm53Flashx,
221
222    // MiMo models
223    /// MiMo V2.6 Pro - Xiaomi's flagship reasoning model with 1M context
224    MiMoV26Pro,
225    /// MiMo V2.6 Flash - Xiaomi's efficient high-volume model with 1M context
226    MiMoV26Flash,
227    /// MiMo V2.6 Pro UltraSpeed - Xiaomi's fastest flagship variant with 1M context
228    MiMoV26ProUltraspeed,
229
230    // Moonshot models
231    /// Kimi K3.1 Pro - Moonshot.ai's 3T-class open-weight flagship post-trained from K3, 1M context
232    MoonshotKimiK31Pro,
233    /// Kimi K3 - Moonshot.ai's 2.8T parameter flagship with Delta Attention, native vision, 1M context
234    MoonshotKimiK3,
235
236    // OpenCode Zen models
237
238    // OpenCode Go models (20 models - https://opencode.ai/docs/go)
239    /// GLM-5.3 - Z.AI flagship for frontier long-horizon coding on OpenCode Go
240    OpenCodeGoGlm53,
241    /// GLM-5.2 - Z.AI flagship model included with OpenCode Go
242    /// GPT-5.6 Luna - OpenAI cost-efficient frontier model on OpenCode Go
243    OpenCodeGoGpt56Luna,
244    /// Kimi K3 - Moonshot flagship 2.8T agentic model on OpenCode Go
245    OpenCodeGoKimiK3,
246    /// MiniMax M3 - Frontier multimodal coding model on OpenCode Go
247    OpenCodeGoMinimaxM3,
248
249    // Qwen models (non-Qwen3 only)
250
251    // Ollama models
252    /// GPT-OSS 20B - Open-weight GPT-OSS 20B model served via Ollama locally
253    OllamaGptOss20b,
254    /// GPT-OSS 20B Cloud - Cloud-hosted GPT-OSS 20B served via Ollama Cloud
255    OllamaGptOss20bCloud,
256    /// GPT-OSS 120B Cloud - Cloud-hosted GPT-OSS 120B served via Ollama Cloud
257    OllamaGptOss120bCloud,
258    /// MiniMax-M3 Cloud - Cloud-hosted MiniMax-M3 model served via Ollama Cloud
259    OllamaMinimaxM3Cloud,
260    /// MiniMax M3.1 Flash Cloud - ultra-efficient multimodal MiniMax M3.1 Flash served via Ollama Cloud
261    OllamaMinimaxM31FlashCloud,
262    /// GLM-5.2 Cloud - Cloud-hosted GLM-5.2 flagship model served via Ollama Cloud
263    /// GLM-5.3 Cloud - Cloud-hosted GLM-5.3 flagship model served via Ollama Cloud
264    OllamaGlm53Cloud,
265    /// GLM-5.3 Flash Cloud - Multimodal GLM-5.3 Flash served via Ollama Cloud
266    OllamaGlm53FlashCloud,
267    /// Mistral Large 4 Cloud - Multimodal Mistral Large 4 served via Ollama Cloud
268    OllamaMistralLarge4Cloud,
269    /// Muse Spark Open Cloud - Multimodal Muse Spark Open served via Ollama Cloud
270    OllamaMuseSparkOpenCloud,
271    /// DeepSeek V4.1 Flash Cloud - Multimodal DeepSeek V4.1 Flash served via Ollama Cloud
272    OllamaDeepSeekV41FlashCloud,
273    /// Kimi K3 Cloud - Moonshot Kimi K3 via Ollama Cloud
274    OllamaKimiK3Cloud,
275    /// Gemma 4 - Google Gemma 4 model served via Ollama
276    OllamaGemma4,
277    /// Laguna XS.2 - Poolside's 33B MoE model (3B activated) for agentic coding via Ollama
278
279    // llama.cpp models
280    /// Gemma 4 26B A4B - Desktop Gemma 4 MoE model served through llama.cpp
281    LlamaCppGemma426bA4b,
282    /// Gemma 4 E4B - Tiny-footprint Gemma 4 model served through llama.cpp
283    LlamaCppGemma4E4b,
284    /// GPT-OSS 20B - OpenAI open-weight model served through llama.cpp
285    LlamaCppGptOss20b,
286
287    // MiniMax models
288    /// MiniMax-M3 - Frontier multimodal coding model with 1M context
289    MinimaxM3,
290    /// MiniMax-M3.1-Flash - Latest ultra-efficient multimodal model via OpenRouter
291    OpenRouterMinimaxM31FlashFree,
292    /// MiniMax-M3.1-Flash - Latest ultra-efficient multimodal model with 1M context and tunable thinking depth
293    MinimaxM31Flash,
294
295    // OpenRouter models
296    /// DeepSeek V4.1 Flash 1017 - Latest flash model via OpenRouter
297    OpenRouterDeepSeekFlash,
298    /// OpenAI gpt-oss-120b - Open-weight 120B reasoning model via OpenRouter
299    OpenRouterOpenAIGptOss120b,
300    /// OpenAI gpt-oss-120b:free - Open-weight 120B reasoning model free tier via OpenRouter
301    OpenRouterOpenAIGptOss120bFree,
302    /// OpenAI gpt-oss-20b - Open-weight 20B deployment via OpenRouter
303    OpenRouterOpenAIGptOss20b,
304    /// OpenAI GPT-6 Astra - OpenAI's flagship model for demanding end-to-end work via OpenRouter
305    OpenRouterOpenAIGpt6Astra,
306    /// GPT-6 Sol - Cost-efficient high-end model in the GPT-6 series via OpenRouter
307    OpenRouterOpenAIGpt6Sol,
308    /// GPT-6 Luna - Fast cost-efficient model in the GPT-6 series via OpenRouter
309    OpenRouterOpenAIGpt6Luna,
310
311    /// Meta Muse Glimmer 30B via OpenRouter
312    OpenRouterMetaMuseGlimmer30b,
313    /// Meta Muse Spark 1.3 via OpenRouter
314    OpenRouterMetaMuseSpark13,
315    /// Meta Muse Spark Open 816B-A42B via OpenRouter
316    OpenRouterMetaMuseSparkOpen,
317    /// Gemini 3.7 Flash - Flash model with 1M context and tunable thinking levels via OpenRouter
318    /// Gemini 3.8 Flash - Most intelligent Flash for long-horizon SWE/agents with 1M context via OpenRouter
319    OpenRouterGoogleGemini38Flash,
320
321    /// Claude Sonnet 5 - Anthropic Claude Sonnet 5 listing
322    OpenRouterAnthropicClaudeSonnet5,
323    /// Mistral Large 3 2512 - Mistral Large 3 2512 model via OpenRouter
324    OpenRouterMistralaiMistralLarge2512,
325    /// DeepSeek V3.1 Nex N1 - Nex AGI DeepSeek V3.1 Nex N1 model via OpenRouter
326    OpenRouterNexAgiDeepseekV31NexN1,
327    /// GLM-5.2 - Z.AI GLM-5.2 flagship model for long-horizon tasks via OpenRouter
328    /// GLM-5.3 Flash - Z.AI efficient multimodal model with hybrid sparse+linear attention via OpenRouter
329    OpenRouterZaiGlm53Flash,
330    /// GLM-5.3 FlashX - Z.AI high-speed Flash variant with faster inference via OpenRouter
331    OpenRouterZaiGlm53Flashx,
332    /// Kimi K3 - Moonshot AI's 2.8T parameter flagship via OpenRouter
333    OpenRouterMoonshotaiKimiK3,
334    /// Grok 4.6 - xAI's flagship reasoning model with reasoning_effort support via OpenRouter
335    OpenRouterXAiGrok46,
336    /// Grok 4.7 - xAI's flagship reasoning model with reasoning_effort support via OpenRouter
337    OpenRouterXAiGrok47,
338    /// MiMo-V2.6-Pro - Xiaomi's flagship agentic model for complex software engineering via OpenRouter
339    OpenRouterXiaomiMimoV26Pro,
340    /// MiMo-V2.6-Flash - Xiaomi's efficient high-volume agentic model via OpenRouter
341    OpenRouterXiaomiMimoV26Flash,
342    /// MiMo-V2.6-Pro-UltraSpeed - Xiaomi's fastest flagship variant via OpenRouter
343    OpenRouterXiaomiMimoV26ProUltraspeed,
344    /// Step 5 Preview - StepFun's flagship agentic model via OpenRouter
345    OpenRouterStepfunStep5Preview,
346
347    // Vercel AI Gateway models (namespaced as `vendor/model` on the gateway)
348    /// Claude Sonnet 5 served through the Vercel AI Gateway
349    VercelAnthropicClaudeSonnet5,
350    /// Claude Opus 5 served through the Vercel AI Gateway
351    VercelAnthropicClaudeOpus5,
352    /// Claude Opus 5.5 served through the Vercel AI Gateway
353    VercelAnthropicClaudeOpus55,
354    /// Claude Haiku 4.5 served through the Vercel AI Gateway
355    VercelAnthropicClaudeHaiku45,
356    /// GPT-5.6 Sol served through the Vercel AI Gateway
357    VercelOpenAiGpt56Sol,
358    /// GPT-6 Astra served through the Vercel AI Gateway
359    VercelOpenAiGpt6Astra,
360    /// GPT-5.6 Luna served through the Vercel AI Gateway
361    VercelOpenAiGpt56Luna,
362    /// GPT-6 Sol served through the Vercel AI Gateway
363    VercelOpenAiGpt6Sol,
364    /// GPT-6 Luna served through the Vercel AI Gateway
365    VercelOpenAiGpt6Luna,
366    /// GPT-5.3 Codex served through the Vercel AI Gateway
367    /// Gemini 3.1 Pro Preview served through the Vercel AI Gateway
368    /// Gemini 3.8 Flash served through the Vercel AI Gateway
369    VercelGoogleGemini38Flash,
370    /// DeepSeek V4.1 Flash served through the Vercel AI Gateway
371    VercelDeepseekFlash,
372    /// Kimi K3 served through the Vercel AI Gateway
373    VercelMoonshotaiKimiK3,
374    /// MiniMax M3 served through the Vercel AI Gateway
375    VercelMinimaxM3,
376    /// Grok 4.7 served through the Vercel AI Gateway
377    VercelSpacexaiGrok47,
378    // xAI models
379    /// Grok 4.6 - xAI's flagship reasoning model with reasoning_effort support (500k context)
380    XaiGrok46,
381    /// Grok 4.7 - xAI's flagship reasoning model with reasoning_effort support (500k context)
382    XaiGrok47,
383
384    /// User-defined model not in the hardcoded catalog.
385    /// Carries the provider key string and model identifier string.
386    Custom(String, String),
387}
vtcode_config/models/model_id.rs

vtcode_config/models/
model_id.rs
