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

vtcode_config/constants/models/
huggingface.rs
