1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
use serde::{Deserialize, Serialize};
mod as_str;
mod capabilities;
mod collection;
mod defaults;
mod description;
mod display;
mod format;
mod openrouter;
mod parse;
mod provider;
mod table;
pub use capabilities::{
ModelCatalogEntry, ModelPricing, catalog_provider_keys, model_catalog_entry, supported_models_for_provider,
};
/// Centralized enum for all supported model identifiers
#[cfg_attr(feature = "schema", derive(schemars::JsonSchema))]
#[derive(Clone, Debug, Default, PartialEq, Eq, Hash, Serialize, Deserialize)]
pub enum ModelId {
// Gemini models
/// Gemini 3.6 Flash - Latest flash model with improved capabilities
/// Gemini 3.7 Flash - Flash model with 1M context and tunable thinking levels
/// Gemini 3.8 Flash - Most intelligent Flash for long-horizon SWE, agents, and enterprise workflows (1M context, 64k output, low/medium/high thinking)
Gemini38Flash,
// OpenAI models
/// GPT-6 Astra - Most capable model for hardest end-to-end work with complex reasoning, coding, computer use, and research
GPT6Astra,
/// GPT-5.6 Sol - Frontier model for complex professional work in the GPT-5.6 family
GPT56Sol,
/// GPT-5.6 Terra - GPT-5.6 model that balances intelligence and cost
GPT56Terra,
/// GPT-5.6 Luna - GPT-5.6 model optimized for cost-sensitive workloads
GPT56Luna,
/// GPT-OSS 20B - OpenAI's open-source 20B parameter model using harmony
OpenAIGptOss20b,
/// GPT-OSS 120B - OpenAI's open-source 120B parameter model using harmony
OpenAIGptOss120b,
// Anthropic models
/// Claude Sonnet 5 - The best combination of speed and intelligence with adaptive thinking on by default
#[default]
ClaudeSonnet5,
/// Claude Fable 5 - Anthropic's most capable widely released model for demanding reasoning and long-horizon agentic work
ClaudeFable5,
/// 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
ClaudeFable51,
/// Claude Opus 5 - Anthropic's newest Opus-tier model with 1M context, thinking on by default
ClaudeOpus5,
/// GitHub Copilot auto model selection
CopilotAuto,
/// GitHub Copilot GPT-5.2 Codex
CopilotGPT52Codex,
/// GitHub Copilot GPT-5.1 Codex Max
CopilotGPT51CodexMax,
/// GitHub Copilot GPT-5.4
CopilotGPT54,
/// GitHub Copilot GPT-5.4 Mini
CopilotGPT54Mini,
// DeepSeek models
/// DeepSeek V4.1 Flash - Latest flash model with improved reasoning and efficiency
DeepSeekFlash,
// Official Meta AI models
/// Meta Muse Spark 1.1 - Official Meta AI Standard-tier reasoning model
MetaMuseSpark11,
/// Meta Muse Spark 1.3 - Official Meta AI flagship Standard-tier reasoning model, tuned for agentic workflows
MetaMuseSpark13,
/// Meta Muse Spark 1.3 Contributor tier - opt-in variant with Meta's discounted data-contribution terms
MetaMuseSpark13Contributor,
// NVIDIA NIM models
/// NVIDIA Nemotron 3 Ultra - NVIDIA's flagship agentic reasoning model via NIM
NvidiaNemotron3Ultra550bA55b,
/// NVIDIA Nemotron 3 Super - Efficient long-context agentic reasoning model via NIM
NvidiaNemotron3Super120bA12b,
/// NVIDIA Nemotron 3 Nano - Efficient reasoning and tool-use model via NIM
NvidiaNemotron3Nano30bA3b,
// Merge Gateway routes
/// Merge Gateway's default route selected by Merge
MergeGatewayDefaultRouting,
/// Anthropic Claude Opus 5 through Merge Gateway
MergeGatewayAnthropicClaudeOpus5,
/// Google Gemini 3.6 Flash through Merge Gateway
/// Google Gemini 3.7 Flash through Merge Gateway
/// DeepSeek V4.1 Flash through Merge Gateway
MergeGatewayDeepseekFlash,
/// xAI Grok 4.6 through Merge Gateway
MergeGatewayXaiGrok46,
/// MiniMax H3 through Merge Gateway
MergeGatewayMinimaxH3,
/// Moonshot Kimi K3 through Merge Gateway
MergeGatewayMoonshotKimiK3,
/// Thinking Machines Inkling through Merge Gateway
MergeGatewayThinkingMachinesInkling,
/// Meta Muse Spark 1.1 through Merge Gateway
MergeGatewayMetaMuseSpark11,
/// Meta Muse Spark 1.3 through Merge Gateway
MergeGatewayMetaMuseSpark13,
/// Z.AI GLM-5.3 Flash through Merge Gateway
MergeGatewayZaiGlm53Flash,
/// OpenAI GPT-5.6 Luna through Merge Gateway
MergeGatewayOpenAIGpt56Luna,
/// OpenAI GPT-5.6 Sol through Merge Gateway
MergeGatewayOpenAIGpt56Sol,
/// OpenAI GPT-5.6 Terra through Merge Gateway
MergeGatewayOpenAIGpt56Terra,
/// Google Gemini 3.8 Flash through Merge Gateway
MergeGatewayGoogleGemini38Flash,
/// Anthropic Claude Fable 5.1 through Merge Gateway
MergeGatewayAnthropicClaudeFable51,
/// OpenAI GPT-6 Astra through Merge Gateway
MergeGatewayOpenAIGpt6Astra,
// Mistral AI models
/// Mistral Large 3 - State-of-the-art open-weight general-purpose multimodal model
MistralLarge3,
// Hugging Face models
/// OpenAI GPT-OSS 20B via Hugging Face router
HuggingFaceOpenAIGptOss20b,
/// OpenAI GPT-OSS 120B via Hugging Face router
HuggingFaceOpenAIGptOss120b,
/// Z.AI GLM-5.2 via Novita inference provider on Hugging Face router
/// Z.AI GLM-5.3 Flash via Together inference provider on Hugging Face router
HuggingFaceGlm53FlashTogether,
/// Z.AI GLM-5.3 via Together inference provider on Hugging Face router
HuggingFaceGlm53Together,
/// Kimi K3 via Together on Hugging Face router
HuggingFaceKimiK3Together,
/// MiniMax M3 via Novita on Hugging Face router
HuggingFaceMinimaxM3Novita,
// StepFun models
/// Step 3.7 Flash - StepFun's flagship multimodal reasoning model with tool calling
StepFun37Flash,
// Evolink gateway models (namespaced as `evolink/<model>`)
/// GPT-5.2 served through the Evolink gateway
/// GPT-5.5 served through the Evolink gateway
/// Gemini 3.1 Pro served through the Evolink gateway (OpenAI SDK format)
EvolinkGemini31Pro,
/// Gemini 3.5 Flash served through the Evolink gateway (OpenAI SDK format)
/// MiniMax-M3 served through the Evolink gateway (OpenAI Chat Completions format)
EvolinkMinimaxM3,
/// Claude Haiku 4.5 served through the Evolink gateway (Anthropic Messages API)
EvolinkClaudeHaiku45,
/// GLM-5.3 - Z.ai flagship coding model with frontier long-horizon agentic performance
ZaiGlm53,
/// GLM-5.3 Flash - Z.ai efficient multimodal model with hybrid sparse+linear attention, 320B total / 18B active, 1M context, native vision
ZaiGlm53Flash,
// MiMo models
/// MiMo V2.5 Pro - Xiaomi's flagship reasoning model with 1M context
MiMoV25Pro,
/// MiMo V2.5 - Xiaomi's omni-modal model with full-modal understanding and 1M context
MiMoV25,
// Moonshot models
/// Kimi K3 - Moonshot.ai's 2.8T parameter flagship with Delta Attention, native vision, 1M context
MoonshotKimiK3,
// OpenCode Zen models
// OpenCode Go models (20 models - https://opencode.ai/docs/go)
/// GLM-5.3 - Z.AI flagship for frontier long-horizon coding on OpenCode Go
OpenCodeGoGlm53,
/// GLM-5.2 - Z.AI flagship model included with OpenCode Go
/// GPT-5.6 Luna - OpenAI cost-efficient frontier model on OpenCode Go
OpenCodeGoGpt56Luna,
/// Kimi K3 - Moonshot flagship 2.8T agentic model on OpenCode Go
OpenCodeGoKimiK3,
/// MiMo-V2.5 - Xiaomi's omnimodal model on OpenCode Go
OpenCodeGoMimoV25,
/// MiMo-V2.5-Pro - Xiaomi's flagship reasoning model on OpenCode Go
OpenCodeGoMimoV25Pro,
/// MiniMax M3 - Frontier multimodal coding model on OpenCode Go
OpenCodeGoMinimaxM3,
/// Muse Spark 1.2 Contributor - Meta long-context reasoning on OpenCode Go (limited regions)
// Qwen models (non-Qwen3 only)
// Ollama models
/// GPT-OSS 20B - Open-weight GPT-OSS 20B model served via Ollama locally
OllamaGptOss20b,
/// GPT-OSS 20B Cloud - Cloud-hosted GPT-OSS 20B served via Ollama Cloud
OllamaGptOss20bCloud,
/// GPT-OSS 120B Cloud - Cloud-hosted GPT-OSS 120B served via Ollama Cloud
OllamaGptOss120bCloud,
/// MiniMax-M3 Cloud - Cloud-hosted MiniMax-M3 model served via Ollama Cloud
OllamaMinimaxM3Cloud,
/// GLM-5.2 Cloud - Cloud-hosted GLM-5.2 flagship model served via Ollama Cloud
/// GLM-5.3 Cloud - Cloud-hosted GLM-5.3 flagship model served via Ollama Cloud
OllamaGlm53Cloud,
/// Kimi K3 Cloud - Moonshot Kimi K3 via Ollama Cloud
OllamaKimiK3Cloud,
/// Gemma 4 - Google Gemma 4 model served via Ollama
OllamaGemma4,
/// Laguna XS.2 - Poolside's 33B MoE model (3B activated) for agentic coding via Ollama
// llama.cpp models
/// Gemma 4 26B A4B - Desktop Gemma 4 MoE model served through llama.cpp
LlamaCppGemma426bA4b,
/// Gemma 4 E4B - Tiny-footprint Gemma 4 model served through llama.cpp
LlamaCppGemma4E4b,
/// GPT-OSS 20B - OpenAI open-weight model served through llama.cpp
LlamaCppGptOss20b,
// MiniMax models
/// MiniMax-M3 - Frontier multimodal coding model with 1M context
MinimaxM3,
// OpenRouter models
/// DeepSeek V4.1 Flash - Latest flash model via OpenRouter
OpenRouterDeepSeekFlash,
/// OpenAI gpt-oss-120b - Open-weight 120B reasoning model via OpenRouter
OpenRouterOpenAIGptOss120b,
/// OpenAI gpt-oss-120b:free - Open-weight 120B reasoning model free tier via OpenRouter
OpenRouterOpenAIGptOss120bFree,
/// OpenAI gpt-oss-20b - Open-weight 20B deployment via OpenRouter
OpenRouterOpenAIGptOss20b,
/// OpenAI GPT-6 Astra - OpenAI's flagship model for demanding end-to-end work via OpenRouter
OpenRouterOpenAIGpt6Astra,
/// Meta Muse Glimmer 30B via OpenRouter
OpenRouterMetaMuseGlimmer30b,
/// Meta Muse Spark 1.2 via OpenRouter
/// Meta Muse Spark 1.3 via OpenRouter
OpenRouterMetaMuseSpark13,
/// Gemini 3.7 Flash - Flash model with 1M context and tunable thinking levels via OpenRouter
/// Gemini 3.8 Flash - Most intelligent Flash for long-horizon SWE/agents with 1M context via OpenRouter
OpenRouterGoogleGemini38Flash,
/// Claude Sonnet 5 - Anthropic Claude Sonnet 5 listing
OpenRouterAnthropicClaudeSonnet5,
/// Mistral Large 3 2512 - Mistral Large 3 2512 model via OpenRouter
OpenRouterMistralaiMistralLarge2512,
/// DeepSeek V3.1 Nex N1 - Nex AGI DeepSeek V3.1 Nex N1 model via OpenRouter
OpenRouterNexAgiDeepseekV31NexN1,
/// GLM-5.2 - Z.AI GLM-5.2 flagship model for long-horizon tasks via OpenRouter
/// GLM-5.3 Flash - Z.AI efficient multimodal model with hybrid sparse+linear attention via OpenRouter
OpenRouterZaiGlm53Flash,
/// Kimi K3 - Moonshot AI's 2.8T parameter flagship via OpenRouter
OpenRouterMoonshotaiKimiK3,
/// Grok 4.6 - xAI's flagship reasoning model with reasoning_effort support via OpenRouter
OpenRouterXAiGrok46,
/// MiMo-V2.5 - Xiaomi's omnimodal agentic model for complex software engineering via OpenRouter
OpenRouterXiaomiMimoV25,
/// MiMo-V2.5-Pro - Xiaomi's flagship agentic model for complex software engineering via OpenRouter
OpenRouterXiaomiMimoV25Pro,
// Vercel AI Gateway models (namespaced as `vendor/model` on the gateway)
/// Claude Sonnet 5 served through the Vercel AI Gateway
VercelAnthropicClaudeSonnet5,
/// Claude Opus 5 served through the Vercel AI Gateway
VercelAnthropicClaudeOpus5,
/// Claude Haiku 4.5 served through the Vercel AI Gateway
VercelAnthropicClaudeHaiku45,
/// GPT-5.6 Sol served through the Vercel AI Gateway
VercelOpenAiGpt56Sol,
/// GPT-6 Astra served through the Vercel AI Gateway
VercelOpenAiGpt6Astra,
/// GPT-5.6 Luna served through the Vercel AI Gateway
VercelOpenAiGpt56Luna,
/// GPT-5.3 Codex served through the Vercel AI Gateway
/// Gemini 3.1 Pro Preview served through the Vercel AI Gateway
/// Gemini 3.8 Flash served through the Vercel AI Gateway
VercelGoogleGemini38Flash,
/// DeepSeek V4.1 Flash served through the Vercel AI Gateway
VercelDeepseekFlash,
/// Kimi K3 served through the Vercel AI Gateway
VercelMoonshotaiKimiK3,
/// MiniMax M3 served through the Vercel AI Gateway
VercelMinimaxM3,
// xAI models
/// Grok 4.6 - xAI's flagship reasoning model with reasoning_effort support (500k context)
XaiGrok46,
/// User-defined model not in the hardcoded catalog.
/// Carries the provider key string and model identifier string.
Custom(String, String),
}