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    DEEPSEEK_V32,
13    OPENAI_GPT_OSS_20B,
14    // Novita inference provider models
15    MINIMAX_M2_5_NOVITA,
16    DEEPSEEK_V32_NOVITA,
17    XIAOMI_MIMO_V2_FLASH_NOVITA,
18    QWEN3_CODER_NEXT_NOVITA,
19    ZAI_GLM_5_NOVITA,
20    ZAI_GLM_5_1_ZAI_ORG,
21    // Moonshot inference provider models
22    KIMI_K2_6_NOVITA,
23    // Together inference provider models
24    QWEN3_5_397B_A17B_TOGETHER,
25    DEEPSEEK_V4_PRO_TOGETHER,
26    STEP_3_5_FLASH,
27];
28
29// Recommended conversational LLMs
30pub const GOOGLE_GEMMA_2_2B_IT: &str = "google/gemma-2-2b-it";
31pub const QWEN3_CODER_480B_A35B_INSTRUCT: &str = "Qwen/Qwen3-Coder-480B-A35B-Instruct";
32pub const OPENAI_GPT_OSS_120B: &str = "openai/gpt-oss-120b:huggingface";
33pub const QWEN3_4B_THINKING_2507: &str = "Qwen/Qwen3-4B-Thinking-2507";
34pub const QWEN25_7B_INSTRUCT_1M: &str = "Qwen/Qwen2.5-7B-Instruct-1M";
35pub const QWEN25_CODER_32B_INSTRUCT: &str = "Qwen/Qwen2.5-Coder-32B-Instruct";
36pub const DEEPSEEK_R1: &str = "deepseek-ai/DeepSeek-R1";
37pub const STEP_3_5_FLASH_BASE: &str = "stepfun-ai/Step-3.5-Flash";
38pub const STEP_3_5_FLASH_PROVIDER: &str = "featherless-ai";
39pub const STEP_3_5_FLASH: &str = "stepfun-ai/Step-3.5-Flash:featherless-ai";
40pub const STEP_3_5_FLASH_LEGACY_FASTEST: &str = "stepfun-ai/Step-3.5-Flash:fastest";
41
42// Additional supported models
43pub const DEEPSEEK_V32: &str = "deepseek-ai/DeepSeek-V3.2:huggingface";
44pub const OPENAI_GPT_OSS_20B: &str = "openai/gpt-oss-20b:huggingface";
45
46pub const MINIMAX_M2_5_NOVITA: &str = "MiniMaxAI/MiniMax-M2.5:novita";
47pub const DEEPSEEK_V32_NOVITA: &str = "deepseek-ai/DeepSeek-V3.2:novita";
48pub const XIAOMI_MIMO_V2_FLASH_NOVITA: &str = "XiaomiMiMo/MiMo-V2-Flash:novita";
49pub const QWEN3_CODER_NEXT_NOVITA: &str = "Qwen/Qwen3-Coder-Next:novita";
50pub const ZAI_GLM_5_NOVITA: &str = "zai-org/GLM-5:novita";
51pub const ZAI_GLM_5_1_ZAI_ORG: &str = "zai-org/GLM-5.1:zai-org";
52pub const KIMI_K2_6_NOVITA: &str = "moonshotai/Kimi-K2.6:novita";
53pub const QWEN3_5_397B_A17B_TOGETHER: &str = "Qwen/Qwen3.5-397B-A17B:together";
54
55// DeepSeek V4 models via HF router
56pub const DEEPSEEK_V4_FLASH_NOVITA: &str = "deepseek-ai/DeepSeek-V4-Flash:novita";
57pub const DEEPSEEK_V4_PRO_TOGETHER: &str = "deepseek-ai/DeepSeek-V4-Pro:together";
58
59pub const REASONING_MODELS: &[&str] = &[
60    // All recommended conversational LLMs support reasoning
61    QWEN3_CODER_480B_A35B_INSTRUCT,
62    OPENAI_GPT_OSS_120B,
63    QWEN3_4B_THINKING_2507,
64    DEEPSEEK_R1,
65    // Additional reasoning models
66    DEEPSEEK_V32,
67    OPENAI_GPT_OSS_20B,
68    DEEPSEEK_V32_NOVITA,
69    MINIMAX_M2_5_NOVITA,
70    XIAOMI_MIMO_V2_FLASH_NOVITA,
71    QWEN3_CODER_NEXT_NOVITA,
72    ZAI_GLM_5_1_ZAI_ORG,
73    QWEN3_5_397B_A17B_TOGETHER,
74    DEEPSEEK_V4_PRO_TOGETHER,
75    DEEPSEEK_V4_FLASH_NOVITA,
76    STEP_3_5_FLASH,
77];
vtcode_config/constants/models/huggingface.rs

vtcode_config/constants/models/
huggingface.rs
