vtcode_config/constants/models/
huggingface.rs

1pub const DEFAULT_MODEL: &str = OPENAI_GPT_OSS_120B;
2pub const SUPPORTED_MODELS: &[&str] = &[
3    // Recommended conversational LLMs from HF docs
4    GOOGLE_GEMMA_2_2B_IT,
5    QWEN3_CODER_480B_A35B_INSTRUCT,
6    OPENAI_GPT_OSS_120B,
7    QWEN3_4B_THINKING_2507,
8    QWEN25_7B_INSTRUCT_1M,
9    QWEN25_CODER_32B_INSTRUCT,
10    DEEPSEEK_R1,
11    // Additional supported models
12    OPENAI_GPT_OSS_20B,
13    // Novita inference provider models
14    MINIMAX_M2_5_NOVITA,
15    XIAOMI_MIMO_V2_FLASH_NOVITA,
16    QWEN3_CODER_NEXT_NOVITA,
17    ZAI_GLM_5_NOVITA,
18    ZAI_GLM_5_1_ZAI_ORG,
19    // Moonshot inference provider models
20    KIMI_K2_6_NOVITA,
21    // Together inference provider models
22    QWEN3_5_397B_A17B_TOGETHER,
23    DEEPSEEK_V4_PRO_TOGETHER,
24    STEP_3_5_FLASH,
25    // DeepInfra inference provider models
26    ZAI_GLM_5_1_DEEPINFRA,
27    // Additional Novita models
28    MINIMAX_M2_7_NOVITA,
29    DEEPSEEK_V4_PRO_NOVITA,
30    NVIDIA_NEMOTRON_3_ULTRA_550B_A55B_NVFP4_TOGETHER,
31];
32
33// Recommended conversational LLMs
34pub const GOOGLE_GEMMA_2_2B_IT: &str = "google/gemma-2-2b-it";
35pub const QWEN3_CODER_480B_A35B_INSTRUCT: &str = "Qwen/Qwen3-Coder-480B-A35B-Instruct";
36pub const OPENAI_GPT_OSS_120B: &str = "openai/gpt-oss-120b:huggingface";
37pub const QWEN3_4B_THINKING_2507: &str = "Qwen/Qwen3-4B-Thinking-2507";
38pub const QWEN25_7B_INSTRUCT_1M: &str = "Qwen/Qwen2.5-7B-Instruct-1M";
39pub const QWEN25_CODER_32B_INSTRUCT: &str = "Qwen/Qwen2.5-Coder-32B-Instruct";
40pub const DEEPSEEK_R1: &str = "deepseek-ai/DeepSeek-R1";
41pub const STEP_3_5_FLASH_BASE: &str = "stepfun-ai/Step-3.5-Flash";
42pub const STEP_3_5_FLASH_PROVIDER: &str = "featherless-ai";
43pub const STEP_3_5_FLASH: &str = "stepfun-ai/Step-3.5-Flash:featherless-ai";
44pub const STEP_3_5_FLASH_LEGACY_FASTEST: &str = "stepfun-ai/Step-3.5-Flash:fastest";
45
46// Additional supported models
47pub const OPENAI_GPT_OSS_20B: &str = "openai/gpt-oss-20b:huggingface";
48
49pub const MINIMAX_M2_5_NOVITA: &str = "MiniMaxAI/MiniMax-M2.5:novita";
50pub const XIAOMI_MIMO_V2_FLASH_NOVITA: &str = "XiaomiMiMo/MiMo-V2-Flash:novita";
51pub const QWEN3_CODER_NEXT_NOVITA: &str = "Qwen/Qwen3-Coder-Next:novita";
52pub const ZAI_GLM_5_NOVITA: &str = "zai-org/GLM-5:novita";
53pub const ZAI_GLM_5_1_ZAI_ORG: &str = "zai-org/GLM-5.1:zai-org";
54pub const KIMI_K2_6_NOVITA: &str = "moonshotai/Kimi-K2.6:novita";
55pub const QWEN3_5_397B_A17B_TOGETHER: &str = "Qwen/Qwen3.5-397B-A17B:together";
56
57// DeepSeek V4 models via HF router
58pub const DEEPSEEK_V4_FLASH_NOVITA: &str = "deepseek-ai/DeepSeek-V4-Flash:novita";
59pub const DEEPSEEK_V4_PRO_TOGETHER: &str = "deepseek-ai/DeepSeek-V4-Pro:together";
60pub const DEEPSEEK_V4_PRO_NOVITA: &str = "deepseek-ai/DeepSeek-V4-Pro:novita";
61
62// NVIDIA models via HF router
63pub const NVIDIA_NEMOTRON_3_ULTRA_550B_A55B_NVFP4_TOGETHER: &str =
64    "nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4:together";
65
66// DeepInfra inference provider models
67pub const ZAI_GLM_5_1_DEEPINFRA: &str = "zai-org/GLM-5.1:deepinfra";
68
69// Additional Novita models
70pub const MINIMAX_M2_7_NOVITA: &str = "MiniMaxAI/MiniMax-M2.7:novita";
71
72pub const REASONING_MODELS: &[&str] = &[
73    // All recommended conversational LLMs support reasoning
74    QWEN3_CODER_480B_A35B_INSTRUCT,
75    OPENAI_GPT_OSS_120B,
76    QWEN3_4B_THINKING_2507,
77    DEEPSEEK_R1,
78    // Additional reasoning models
79    OPENAI_GPT_OSS_20B,
80    MINIMAX_M2_5_NOVITA,
81    XIAOMI_MIMO_V2_FLASH_NOVITA,
82    QWEN3_CODER_NEXT_NOVITA,
83    ZAI_GLM_5_1_ZAI_ORG,
84    ZAI_GLM_5_1_DEEPINFRA,
85    MINIMAX_M2_7_NOVITA,
86    QWEN3_5_397B_A17B_TOGETHER,
87    DEEPSEEK_V4_PRO_TOGETHER,
88    DEEPSEEK_V4_PRO_NOVITA,
89    DEEPSEEK_V4_FLASH_NOVITA,
90    STEP_3_5_FLASH,
91    NVIDIA_NEMOTRON_3_ULTRA_550B_A55B_NVFP4_TOGETHER,
92];
vtcode_config/constants/models/huggingface.rs

vtcode_config/constants/models/
huggingface.rs
