diff --git a/.cspell-wordlist.txt b/.cspell-wordlist.txt
index 3742847485..4cae5e07bb 100644
--- a/.cspell-wordlist.txt
+++ b/.cspell-wordlist.txt
@@ -52,4 +52,6 @@ MPNET
 QINT
 FNUZ
 wordlist
+coreml
+mobilenetv
 jitpack
diff --git a/apps/llm/app/llm/index.tsx b/apps/llm/app/llm/index.tsx
index f59e9e6894..2a647b4adb 100644
--- a/apps/llm/app/llm/index.tsx
+++ b/apps/llm/app/llm/index.tsx
@@ -12,12 +12,7 @@ import {
 } from 'react-native';
 import SendIcon from '../../assets/icons/send_icon.svg';
 import Spinner from 'react-native-loading-spinner-overlay';
-import {
-  LLAMA3_2_1B_QLORA,
-  LLAMA3_2_TOKENIZER,
-  LLAMA3_2_TOKENIZER_CONFIG,
-  useLLM,
-} from 'react-native-executorch';
+import { LLAMA3_2_1B_QLORA, useLLM } from 'react-native-executorch';
 import PauseIcon from '../../assets/icons/pause_icon.svg';
 import ColorPalette from '../../colors';
 import Messages from '../../components/Messages';
@@ -35,11 +30,7 @@ function LLMScreen() {
   const textInputRef = useRef<TextInput>(null);
   const { setGlobalGenerating } = useContext(GeneratingContext);
 
-  const llm = useLLM({
-    modelSource: LLAMA3_2_1B_QLORA,
-    tokenizerSource: LLAMA3_2_TOKENIZER,
-    tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
-  });
+  const llm = useLLM(LLAMA3_2_1B_QLORA);
 
   useEffect(() => {
     if (llm.error) {
diff --git a/apps/llm/app/llm_structured_output/index.tsx b/apps/llm/app/llm_structured_output/index.tsx
index dd6a7275be..4bfa4d7163 100644
--- a/apps/llm/app/llm_structured_output/index.tsx
+++ b/apps/llm/app/llm_structured_output/index.tsx
@@ -13,8 +13,6 @@ import {
 import SendIcon from '../../assets/icons/send_icon.svg';
 import Spinner from 'react-native-loading-spinner-overlay';
 import {
-  QWEN3_TOKENIZER,
-  QWEN3_TOKENIZER_CONFIG,
   useLLM,
   fixAndValidateStructuredOutput,
   getStructuredOutputPrompt,
@@ -75,12 +73,8 @@ function LLMScreen() {
   const textInputRef = useRef<TextInput>(null);
   const { setGlobalGenerating } = useContext(GeneratingContext);
 
-  const llm = useLLM({
-    // try out 4B model it this one struggles with following structured output
-    modelSource: QWEN3_1_7B_QUANTIZED,
-    tokenizerSource: QWEN3_TOKENIZER,
-    tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
-  });
+  // try out 4B model if this one struggles with following structured output
+  const llm = useLLM(QWEN3_1_7B_QUANTIZED);
 
   useEffect(() => {
     setGlobalGenerating(llm.isGenerating);
diff --git a/apps/llm/app/llm_tool_calling/index.tsx b/apps/llm/app/llm_tool_calling/index.tsx
index 5d5995f882..2caac8065b 100644
--- a/apps/llm/app/llm_tool_calling/index.tsx
+++ b/apps/llm/app/llm_tool_calling/index.tsx
@@ -15,8 +15,6 @@ import SendIcon from '../../assets/icons/send_icon.svg';
 import Spinner from 'react-native-loading-spinner-overlay';
 import {
   HAMMER2_1_1_5B,
-  HAMMER2_1_TOKENIZER,
-  HAMMER2_1_TOKENIZER_CONFIG,
   useLLM,
   DEFAULT_SYSTEM_PROMPT,
 } from 'react-native-executorch';
@@ -41,11 +39,7 @@ function LLMToolCallingScreen() {
   const textInputRef = useRef<TextInput>(null);
   const { setGlobalGenerating } = useContext(GeneratingContext);
 
-  const llm = useLLM({
-    modelSource: HAMMER2_1_1_5B,
-    tokenizerSource: HAMMER2_1_TOKENIZER,
-    tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
-  });
+  const llm = useLLM(HAMMER2_1_1_5B);
 
   useEffect(() => {
     setGlobalGenerating(llm.isGenerating);
diff --git a/apps/llm/app/voice_chat/index.tsx b/apps/llm/app/voice_chat/index.tsx
index c4baaa9870..f86657281c 100644
--- a/apps/llm/app/voice_chat/index.tsx
+++ b/apps/llm/app/voice_chat/index.tsx
@@ -16,8 +16,6 @@ import {
   useSpeechToText,
   useLLM,
   QWEN3_0_6B_QUANTIZED,
-  QWEN3_TOKENIZER,
-  QWEN3_TOKENIZER_CONFIG,
 } from 'react-native-executorch';
 import PauseIcon from '../../assets/icons/pause_icon.svg';
 import MicIcon from '../../assets/icons/mic_icon.svg';
@@ -68,11 +66,7 @@ function VoiceChatScreen() {
   const messageRecorded = useRef<boolean>(false);
   const { setGlobalGenerating } = useContext(GeneratingContext);
 
-  const llm = useLLM({
-    modelSource: QWEN3_0_6B_QUANTIZED,
-    tokenizerSource: QWEN3_TOKENIZER,
-    tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
-  });
+  const llm = useLLM(QWEN3_0_6B_QUANTIZED);
   const speechToText = useSpeechToText({
     modelName: 'moonshine',
     windowSize: 3,
diff --git a/apps/text-embeddings/app/text-embeddings/index.tsx b/apps/text-embeddings/app/text-embeddings/index.tsx
index 618a40bcc9..ddd8f7be9d 100644
--- a/apps/text-embeddings/app/text-embeddings/index.tsx
+++ b/apps/text-embeddings/app/text-embeddings/index.tsx
@@ -11,11 +11,7 @@ import {
   Platform,
 } from 'react-native';
 import { Ionicons } from '@expo/vector-icons';
-import {
-  useTextEmbeddings,
-  ALL_MINILM_L6_V2,
-  ALL_MINILM_L6_V2_TOKENIZER,
-} from 'react-native-executorch';
+import { useTextEmbeddings, ALL_MINILM_L6_V2 } from 'react-native-executorch';
 import { useIsFocused } from '@react-navigation/native';
 
 export default function TextEmbeddingsScreenWrapper() {
@@ -25,11 +21,7 @@ export default function TextEmbeddingsScreenWrapper() {
 }
 
 function TextEmbeddingsScreen() {
-  const model = useTextEmbeddings({
-    modelSource: ALL_MINILM_L6_V2,
-    tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER,
-    meanPooling: true,
-  });
+  const model = useTextEmbeddings(ALL_MINILM_L6_V2);
 
   const [inputSentence, setInputSentence] = useState('');
   const [sentencesWithEmbeddings, setSentencesWithEmbeddings] = useState<
diff --git a/packages/react-native-executorch/src/constants/modelUrls.ts b/packages/react-native-executorch/src/constants/modelUrls.ts
index b7e83f4798..14944fe704 100644
--- a/packages/react-native-executorch/src/constants/modelUrls.ts
+++ b/packages/react-native-executorch/src/constants/modelUrls.ts
@@ -1,196 +1,388 @@
 import { Platform } from 'react-native';
 
-// LLM's
-
-// LLAMA 3.2
-export const LLAMA3_2_3B =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-3B/original/llama3_2_3B_bf16.pte';
-export const LLAMA3_2_3B_QLORA =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-3B/QLoRA/llama3_2-3B_qat_lora.pte';
-export const LLAMA3_2_3B_SPINQUANT =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-3B/spinquant/llama3_2_3B_spinquant.pte';
-export const LLAMA3_2_1B =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-1B/original/llama3_2_bf16.pte';
-export const LLAMA3_2_1B_QLORA =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-1B/QLoRA/llama3_2_qat_lora.pte';
-export const LLAMA3_2_1B_SPINQUANT =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/llama-3.2-1B/spinquant/llama3_2_spinquant.pte';
-export const LLAMA3_2_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/tokenizer.json';
-export const LLAMA3_2_TOKENIZER_CONFIG =
-  'https://huggingface.co/software-mansion/react-native-executorch-llama-3.2/resolve/v0.4.0/tokenizer_config.json';
-
-// QWEN 3
-export const QWEN3_0_6B =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-0.6B/original/qwen3_0_6b_bf16.pte';
-export const QWEN3_0_6B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-0.6B/quantized/qwen3_0_6b_8da4w.pte';
-export const QWEN3_1_7B =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-1.7B/original/qwen3_1_7b_bf16.pte';
-export const QWEN3_1_7B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-1.7B/quantized/qwen3_1_7b_8da4w.pte';
-export const QWEN3_4B =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-4B/original/qwen3_4b_bf16.pte';
-export const QWEN3_4B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/qwen-3-4B/quantized/qwen3_4b_8da4w.pte';
-export const QWEN3_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/tokenizer.json';
-export const QWEN3_TOKENIZER_CONFIG =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-3/resolve/v0.4.0/tokenizer_config.json';
-
-// HAMMER 2.1
-export const HAMMER2_1_0_5B =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-0.5B/original/hammer2_1_0_5B_bf16.pte';
-export const HAMMER2_1_0_5B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-0.5B/quantized/hammer2_1_0_5B_8da4w.pte';
-export const HAMMER2_1_1_5B =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-1.5B/original/hammer2_1_1_5B_bf16.pte';
-export const HAMMER2_1_1_5B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-1.5B/quantized/hammer2_1_1_5B_8da4w.pte';
-export const HAMMER2_1_3B =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-3B/original/hammer2_1_3B_bf16.pte';
-export const HAMMER2_1_3B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/hammer-2.1-3B/quantized/hammer2_1_3B_8da4w.pte';
-export const HAMMER2_1_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/tokenizer.json';
-export const HAMMER2_1_TOKENIZER_CONFIG =
-  'https://huggingface.co/software-mansion/react-native-executorch-hammer-2.1/resolve/v0.4.0/tokenizer_config.json';
-
-// SMOLLM2
-export const SMOLLM2_1_135M =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-135M/original/smolLm2_135M_bf16.pte';
-export const SMOLLM2_1_135M_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-135M/quantized/smolLm2_135M_8da4w.pte';
-export const SMOLLM2_1_360M =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-360M/original/smolLm2_360M_bf16.pte';
-export const SMOLLM2_1_360M_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-360M/quantized/smolLm2_360M_8da4w.pte';
-export const SMOLLM2_1_1_7B =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-1.7B/original/smolLm2_1_7B_bf16.pte';
-export const SMOLLM2_1_1_7B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/smolLm-2-1.7B/quantized/smolLm2_1_7B_8da4w.pte';
-export const SMOLLM2_1_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/tokenizer.json';
-export const SMOLLM2_1_TOKENIZER_CONFIG =
-  'https://huggingface.co/software-mansion/react-native-executorch-smolLm-2/resolve/v0.4.0/tokenizer_config.json';
-
-// QWEN 2.5
-export const QWEN2_5_0_5B =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-0.5B/original/qwen2_5_0_5b_bf16.pte';
-export const QWEN2_5_0_5B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-0.5B/quantized/qwen2_5_0_5b_8da4w.pte';
-export const QWEN2_5_1_5B =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-1.5B/original/qwen2_5_1_5b_bf16.pte';
-export const QWEN2_5_1_5B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-1.5B/quantized/qwen2_5_1_5b_8da4w.pte';
-export const QWEN2_5_3B =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-3B/original/qwen2_5_3b_bf16.pte';
-export const QWEN2_5_3B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/qwen-2.5-3B/quantized/qwen2_5_3b_8da4w.pte';
-export const QWEN2_5_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/tokenizer.json';
-export const QWEN2_5_TOKENIZER_CONFIG =
-  'https://huggingface.co/software-mansion/react-native-executorch-qwen-2.5/resolve/v0.4.0/tokenizer_config.json';
-
-// PHI 4
-export const PHI_4_MINI_4B =
-  'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/original/phi-4-mini_bf16.pte';
-export const PHI_4_MINI_4B_QUANTIZED =
-  'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/quantized/phi-4-mini_8da4w.pte';
-export const PHI_4_MINI_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/tokenizer.json';
-export const PHI_4_MINI_TOKENIZER_CONFIG =
-  'https://huggingface.co/software-mansion/react-native-executorch-phi-4-mini/resolve/v0.4.0/tokenizer_config.json';
+const BASE_URL_PREFIX =
+  'https://huggingface.co/software-mansion/react-native-executorch';
+const VERSION_TAG = 'resolve/v0.4.0';
+const NEXT_DEV_VERSION_TAG = 'resolve/v0.5.0';
+
+const isIOS = Platform.OS === `ios`;
+
+// LLM`s
+
+const LLAMA3_2_MODEL_NAME = 'llama-3.2';
+const LLAMA3_2_3B_MODEL = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/llama-3.2-3B/original/llama3_2_3B_bf16.pte`;
+const LLAMA3_2_3B_QLORA_MODEL = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/llama-3.2-3B/QLoRA/llama3_2-3B_qat_lora.pte`;
+const LLAMA3_2_3B_SPINQUANT_MODEL = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/llama-3.2-3B/spinquant/llama3_2_3B_spinquant.pte`;
+const LLAMA3_2_1B_MODEL = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/llama-3.2-1B/original/llama3_2_bf16.pte`;
+const LLAMA3_2_1B_QLORA_MODEL = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/llama-3.2-1B/QLoRA/llama3_2_qat_lora.pte`;
+const LLAMA3_2_1B_SPINQUANT_MODEL = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/llama-3.2-1B/spinquant/llama3_2_spinquant.pte`;
+const LLAMA3_2_TOKENIZER = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+const LLAMA3_2_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-${LLAMA3_2_MODEL_NAME}/${VERSION_TAG}/tokenizer_config.json`;
+
+export const LLAMA3_2_3B = {
+  modelSource: LLAMA3_2_3B_MODEL,
+  tokenizerSource: LLAMA3_2_TOKENIZER,
+  tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
+};
+
+export const LLAMA3_2_3B_QLORA = {
+  modelSource: LLAMA3_2_3B_QLORA_MODEL,
+  tokenizerSource: LLAMA3_2_TOKENIZER,
+  tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
+};
+
+export const LLAMA3_2_3B_SPINQUANT = {
+  modelSource: LLAMA3_2_3B_SPINQUANT_MODEL,
+  tokenizerSource: LLAMA3_2_TOKENIZER,
+  tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
+};
+
+export const LLAMA3_2_1B = {
+  modelSource: LLAMA3_2_1B_MODEL,
+  tokenizerSource: LLAMA3_2_TOKENIZER,
+  tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
+};
+
+export const LLAMA3_2_1B_QLORA = {
+  modelSource: LLAMA3_2_1B_QLORA_MODEL,
+  tokenizerSource: LLAMA3_2_TOKENIZER,
+  tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
+};
+
+export const LLAMA3_2_1B_SPINQUANT = {
+  modelSource: LLAMA3_2_1B_SPINQUANT_MODEL,
+  tokenizerSource: LLAMA3_2_TOKENIZER,
+  tokenizerConfigSource: LLAMA3_2_TOKENIZER_CONFIG,
+};
+
+const QWEN3_MODEL_NAME = 'qwen-3';
+const QWEN3_0_6B_MODEL = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/qwen-3-0.6B/original/qwen3_0_6b_bf16.pte`;
+const QWEN3_0_6B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/qwen-3-0.6B/quantized/qwen3_0_6b_8da4w.pte`;
+const QWEN3_1_7B_MODEL = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/qwen-3-1.7B/original/qwen3_1_7b_bf16.pte`;
+const QWEN3_1_7B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/qwen-3-1.7B/quantized/qwen3_1_7b_8da4w.pte`;
+const QWEN3_4B_MODEL = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/qwen-3-4B/original/qwen3_4b_bf16.pte`;
+const QWEN3_4B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/qwen-3-4B/quantized/qwen3_4b_8da4w.pte`;
+const QWEN3_TOKENIZER = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+const QWEN3_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-${QWEN3_MODEL_NAME}/${VERSION_TAG}/tokenizer_config.json`;
+
+export const QWEN3_0_6B = {
+  modelSource: QWEN3_0_6B_MODEL,
+  tokenizerSource: QWEN3_TOKENIZER,
+  tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
+};
+
+export const QWEN3_0_6B_QUANTIZED = {
+  modelSource: QWEN3_0_6B_QUANTIZED_MODEL,
+  tokenizerSource: QWEN3_TOKENIZER,
+  tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
+};
+
+export const QWEN3_1_7B = {
+  modelSource: QWEN3_1_7B_MODEL,
+  tokenizerSource: QWEN3_TOKENIZER,
+  tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
+};
+
+export const QWEN3_1_7B_QUANTIZED = {
+  modelSource: QWEN3_1_7B_QUANTIZED_MODEL,
+  tokenizerSource: QWEN3_TOKENIZER,
+  tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
+};
+
+export const QWEN3_4B = {
+  modelSource: QWEN3_4B_MODEL,
+  tokenizerSource: QWEN3_TOKENIZER,
+  tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
+};
+
+export const QWEN3_4B_QUANTIZED = {
+  modelSource: QWEN3_4B_QUANTIZED_MODEL,
+  tokenizerSource: QWEN3_TOKENIZER,
+  tokenizerConfigSource: QWEN3_TOKENIZER_CONFIG,
+};
+
+const HAMMER2_1_MODEL_NAME = 'hammer-2.1';
+const HAMMER2_1_0_5B_MODEL = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/hammer-2.1-0.5B/original/hammer2_1_0_5B_bf16.pte`;
+const HAMMER2_1_0_5B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/hammer-2.1-0.5B/quantized/hammer2_1_0_5B_8da4w.pte`;
+const HAMMER2_1_1_5B_MODEL = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/hammer-2.1-1.5B/original/hammer2_1_1_5B_bf16.pte`;
+const HAMMER2_1_1_5B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/hammer-2.1-1.5B/quantized/hammer2_1_1_5B_8da4w.pte`;
+const HAMMER2_1_3B_MODEL = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/hammer-2.1-3B/original/hammer2_1_3B_bf16.pte`;
+const HAMMER2_1_3B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/hammer-2.1-3B/quantized/hammer2_1_3B_8da4w.pte`;
+const HAMMER2_1_TOKENIZER = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+const HAMMER2_1_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-${HAMMER2_1_MODEL_NAME}/${VERSION_TAG}/tokenizer_config.json`;
+
+export const HAMMER2_1_0_5B = {
+  modelSource: HAMMER2_1_0_5B_MODEL,
+  tokenizerSource: HAMMER2_1_TOKENIZER,
+  tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
+};
+
+export const HAMMER2_1_0_5B_QUANTIZED = {
+  modelSource: HAMMER2_1_0_5B_QUANTIZED_MODEL,
+  tokenizerSource: HAMMER2_1_TOKENIZER,
+  tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
+};
+
+export const HAMMER2_1_1_5B = {
+  modelSource: HAMMER2_1_1_5B_MODEL,
+  tokenizerSource: HAMMER2_1_TOKENIZER,
+  tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
+};
+
+export const HAMMER2_1_1_5B_QUANTIZED = {
+  modelSource: HAMMER2_1_1_5B_QUANTIZED_MODEL,
+  tokenizerSource: HAMMER2_1_TOKENIZER,
+  tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
+};
+
+export const HAMMER2_1_3B = {
+  modelSource: HAMMER2_1_3B_MODEL,
+  tokenizerSource: HAMMER2_1_TOKENIZER,
+  tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
+};
+
+export const HAMMER2_1_3B_QUANTIZED = {
+  modelSource: HAMMER2_1_3B_QUANTIZED_MODEL,
+  tokenizerSource: HAMMER2_1_TOKENIZER,
+  tokenizerConfigSource: HAMMER2_1_TOKENIZER_CONFIG,
+};
+
+const SMOLLM2_MODEL_NAME = 'smolLm-2';
+const SMOLLM2_1_135M_MODEL = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/smolLm-2-135M/original/smolLm2_135M_bf16.pte`;
+const SMOLLM2_1_135M_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/smolLm-2-135M/quantized/smolLm2_135M_8da4w.pte`;
+const SMOLLM2_1_360M_MODEL = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/smolLm-2-360M/original/smolLm2_360M_bf16.pte`;
+const SMOLLM2_1_360M_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/smolLm-2-360M/quantized/smolLm2_360M_8da4w.pte`;
+const SMOLLM2_1_1_7B_MODEL = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/smolLm-2-1.7B/original/smolLm2_1_7B_bf16.pte`;
+const SMOLLM2_1_1_7B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/smolLm-2-1.7B/quantized/smolLm2_1_7B_8da4w.pte`;
+const SMOLLM2_1_TOKENIZER = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+const SMOLLM2_1_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-${SMOLLM2_MODEL_NAME}/${VERSION_TAG}/tokenizer_config.json`;
+
+export const SMOLLM2_1_135M = {
+  modelSource: SMOLLM2_1_135M_MODEL,
+  tokenizerSource: SMOLLM2_1_TOKENIZER,
+  tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG,
+};
+
+export const SMOLLM2_1_135M_QUANTIZED = {
+  modelSource: SMOLLM2_1_135M_QUANTIZED_MODEL,
+  tokenizerSource: SMOLLM2_1_TOKENIZER,
+  tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG,
+};
+
+export const SMOLLM2_1_360M = {
+  modelSource: SMOLLM2_1_360M_MODEL,
+  tokenizerSource: SMOLLM2_1_TOKENIZER,
+  tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG,
+};
+
+export const SMOLLM2_1_360M_QUANTIZED = {
+  modelSource: SMOLLM2_1_360M_QUANTIZED_MODEL,
+  tokenizerSource: SMOLLM2_1_TOKENIZER,
+  tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG,
+};
+
+export const SMOLLM2_1_1_7B = {
+  modelSource: SMOLLM2_1_1_7B_MODEL,
+  tokenizerSource: SMOLLM2_1_TOKENIZER,
+  tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG,
+};
+
+export const SMOLLM2_1_1_7B_QUANTIZED = {
+  modelSource: SMOLLM2_1_1_7B_QUANTIZED_MODEL,
+  tokenizerSource: SMOLLM2_1_TOKENIZER,
+  tokenizerConfigSource: SMOLLM2_1_TOKENIZER_CONFIG,
+};
+
+const QWEN2_5_MODEL_NAME = 'qwen-2.5';
+const QWEN2_5_0_5B_MODEL = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/qwen-2.5-0.5B/original/qwen2_5_0_5b_bf16.pte`;
+const QWEN2_5_0_5B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/qwen-2.5-0.5B/quantized/qwen2_5_0_5b_8da4w.pte`;
+const QWEN2_5_1_5B_MODEL = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/qwen-2.5-1.5B/original/qwen2_5_1_5b_bf16.pte`;
+const QWEN2_5_1_5B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/qwen-2.5-1.5B/quantized/qwen2_5_1_5b_8da4w.pte`;
+const QWEN2_5_3B_MODEL = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/qwen-2.5-3B/original/qwen2_5_3b_bf16.pte`;
+const QWEN2_5_3B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/qwen-2.5-3B/quantized/qwen2_5_3b_8da4w.pte`;
+const QWEN2_5_TOKENIZER = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+const QWEN2_5_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-${QWEN2_5_MODEL_NAME}/${VERSION_TAG}/tokenizer_config.json`;
+
+export const QWEN2_5_0_5B = {
+  modelSource: QWEN2_5_0_5B_MODEL,
+  tokenizerSource: QWEN2_5_TOKENIZER,
+  tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG,
+};
+
+export const QWEN2_5_0_5B_QUANTIZED = {
+  modelSource: QWEN2_5_0_5B_QUANTIZED_MODEL,
+  tokenizerSource: QWEN2_5_TOKENIZER,
+  tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG,
+};
+
+export const QWEN2_5_1_5B = {
+  modelSource: QWEN2_5_1_5B_MODEL,
+  tokenizerSource: QWEN2_5_TOKENIZER,
+  tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG,
+};
+
+export const QWEN2_5_1_5B_QUANTIZED = {
+  modelSource: QWEN2_5_1_5B_QUANTIZED_MODEL,
+  tokenizerSource: QWEN2_5_TOKENIZER,
+  tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG,
+};
+
+export const QWEN2_5_3B = {
+  modelSource: QWEN2_5_3B_MODEL,
+  tokenizerSource: QWEN2_5_TOKENIZER,
+  tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG,
+};
+
+export const QWEN2_5_3B_QUANTIZED = {
+  modelSource: QWEN2_5_3B_QUANTIZED_MODEL,
+  tokenizerSource: QWEN2_5_TOKENIZER,
+  tokenizerConfigSource: QWEN2_5_TOKENIZER_CONFIG,
+};
+
+const PHI_4_MINI_MODEL_NAME = 'phi-4-mini';
+const PHI_4_MINI_4B_MODEL = `${BASE_URL_PREFIX}-${PHI_4_MINI_MODEL_NAME}/${VERSION_TAG}/original/phi-4-mini_bf16.pte`;
+const PHI_4_MINI_4B_QUANTIZED_MODEL = `${BASE_URL_PREFIX}-${PHI_4_MINI_MODEL_NAME}/${VERSION_TAG}/quantized/phi-4-mini_8da4w.pte`;
+const PHI_4_MINI_TOKENIZER = `${BASE_URL_PREFIX}-${PHI_4_MINI_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+const PHI_4_MINI_TOKENIZER_CONFIG = `${BASE_URL_PREFIX}-${PHI_4_MINI_MODEL_NAME}/${VERSION_TAG}/tokenizer_config.json`;
+
+export const PHI_4_MINI_4B = {
+  modelSource: PHI_4_MINI_4B_MODEL,
+  tokenizerSource: PHI_4_MINI_TOKENIZER,
+  tokenizerConfigSource: PHI_4_MINI_TOKENIZER_CONFIG,
+};
+
+export const PHI_4_MINI_4B_QUANTIZED = {
+  modelSource: PHI_4_MINI_4B_QUANTIZED_MODEL,
+  tokenizerSource: PHI_4_MINI_TOKENIZER,
+  tokenizerConfigSource: PHI_4_MINI_TOKENIZER,
+};
 
 // Classification
-export const EFFICIENTNET_V2_S =
-  Platform.OS === 'ios'
-    ? 'https://huggingface.co/software-mansion/react-native-executorch-efficientnet-v2-s/resolve/v0.4.0/coreml/efficientnet_v2_s_coreml_all.pte'
-    : 'https://huggingface.co/software-mansion/react-native-executorch-efficientnet-v2-s/resolve/v0.4.0/xnnpack/efficientnet_v2_s_xnnpack.pte';
+const EFFICIENTNET_V2_S_MODEL_NAME = 'efficientnet-v2-s';
+export const EFFICIENTNET_V2_S = isIOS
+  ? `${BASE_URL_PREFIX}-${EFFICIENTNET_V2_S_MODEL_NAME}/${VERSION_TAG}/coreml/efficientnet_v2_s_coreml_all.pte`
+  : `${BASE_URL_PREFIX}-${EFFICIENTNET_V2_S_MODEL_NAME}/${VERSION_TAG}/xnnpack/efficientnet_v2_s_xnnpack.pte`;
 
 // Object detection
-export const SSDLITE_320_MOBILENET_V3_LARGE =
-  'https://huggingface.co/software-mansion/react-native-executorch-ssdlite320-mobilenet-v3-large/resolve/v0.4.0/ssdlite320-mobilenetv3-large.pte';
+const SSDLITE_320_MOBILENET_V3_LARGE_MODEL_NAME =
+  'ssdlite320-mobilenet-v3-large';
+export const SSDLITE_320_MOBILENET_V3_LARGE = `${BASE_URL_PREFIX}-${SSDLITE_320_MOBILENET_V3_LARGE_MODEL_NAME}/${VERSION_TAG}/ssdlite320-mobilenetv3-large.pte`;
 
 // Style transfer
-export const STYLE_TRANSFER_CANDY =
-  Platform.OS === 'ios'
-    ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-candy/resolve/v0.4.0/coreml/style_transfer_candy_coreml.pte'
-    : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-candy/resolve/v0.4.0/xnnpack/style_transfer_candy_xnnpack.pte';
-export const STYLE_TRANSFER_MOSAIC =
-  Platform.OS === 'ios'
-    ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-mosaic/resolve/v0.4.0/coreml/style_transfer_mosaic_coreml.pte'
-    : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-mosaic/resolve/v0.4.0/xnnpack/style_transfer_mosaic_xnnpack.pte';
-export const STYLE_TRANSFER_RAIN_PRINCESS =
-  Platform.OS === 'ios'
-    ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-rain-princess/resolve/v0.4.0/coreml/style_transfer_rain_princess_coreml.pte'
-    : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-rain-princess/resolve/v0.4.0/xnnpack/style_transfer_rain_princess_xnnpack.pte';
-export const STYLE_TRANSFER_UDNIE =
-  Platform.OS === 'ios'
-    ? 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-udnie/resolve/v0.4.0/coreml/style_transfer_udnie_coreml.pte'
-    : 'https://huggingface.co/software-mansion/react-native-executorch-style-transfer-udnie/resolve/v0.4.0/xnnpack/style_transfer_udnie_xnnpack.pte';
+const STYLE_TRANSFER_CANDY_MODEL_NAME = 'style-transfer-candy';
+export const STYLE_TRANSFER_CANDY = isIOS
+  ? `${BASE_URL_PREFIX}-${STYLE_TRANSFER_CANDY_MODEL_NAME}/${VERSION_TAG}/coreml/style_transfer_candy_coreml.pte`
+  : `${BASE_URL_PREFIX}-${STYLE_TRANSFER_CANDY_MODEL_NAME}/${VERSION_TAG}/xnnpack/style_transfer_candy_xnnpack.pte`;
+
+const STYLE_TRANSFER_MOSAIC_MODEL_NAME = 'style-transfer-mosaic';
+export const STYLE_TRANSFER_MOSAIC = isIOS
+  ? `${BASE_URL_PREFIX}-${STYLE_TRANSFER_MOSAIC_MODEL_NAME}/${VERSION_TAG}/coreml/style_transfer_mosaic_coreml.pte`
+  : `${BASE_URL_PREFIX}-${STYLE_TRANSFER_MOSAIC_MODEL_NAME}/${VERSION_TAG}/xnnpack/style_transfer_mosaic_xnnpack.pte`;
+
+const STYLE_TRANSFER_RAIN_PRINCESS_MODEL_NAME = 'style-transfer-rain-princess';
+export const STYLE_TRANSFER_RAIN_PRINCESS = isIOS
+  ? `${BASE_URL_PREFIX}-${STYLE_TRANSFER_RAIN_PRINCESS_MODEL_NAME}/${VERSION_TAG}/coreml/style_transfer_rain_princess_coreml.pte`
+  : `${BASE_URL_PREFIX}-${STYLE_TRANSFER_RAIN_PRINCESS_MODEL_NAME}/${VERSION_TAG}/xnnpack/style_transfer_rain_princess_xnnpack.pte`;
+
+const STYLE_TRANSFER_UDNIE_MODEL_NAME = 'style-transfer-udnie';
+export const STYLE_TRANSFER_UDNIE = isIOS
+  ? `${BASE_URL_PREFIX}-${STYLE_TRANSFER_UDNIE_MODEL_NAME}/${VERSION_TAG}/coreml/style_transfer_udnie_coreml.pte`
+  : `${BASE_URL_PREFIX}-${STYLE_TRANSFER_UDNIE_MODEL_NAME}/${VERSION_TAG}/xnnpack/style_transfer_udnie_xnnpack.pte`;
 
 // S2T
-export const MOONSHINE_TINY_DECODER =
-  'https://huggingface.co/software-mansion/react-native-executorch-moonshine-tiny/resolve/v0.4.0/xnnpack/moonshine_tiny_xnnpack_decoder.pte';
-export const MOONSHINE_TINY_ENCODER =
-  'https://huggingface.co/software-mansion/react-native-executorch-moonshine-tiny/resolve/v0.4.0/xnnpack/moonshine_tiny_xnnpack_encoder.pte';
-export const MOONSHINE_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-moonshine-tiny/resolve/v0.4.0/moonshine_tiny_tokenizer.json';
-export const WHISPER_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny.en/resolve/v0.4.0/whisper_tokenizer.json';
-export const WHISPER_TINY_DECODER =
-  'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny.en/resolve/v0.4.0/xnnpack/whisper_tiny_en_xnnpack_decoder.pte';
-export const WHISPER_TINY_ENCODER =
-  'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny.en/resolve/v0.4.0/xnnpack/whisper_tiny_en_xnnpack_encoder.pte';
-export const WHISPER_TINY_MULTILINGUAL_ENCODER =
-  'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny/resolve/v0.4.0/xnnpack/xnnpack_whisper_encoder.pte';
-export const WHISPER_TINY_MULTILINGUAL_DECODER =
-  'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny/resolve/v0.4.0/xnnpack/xnnpack_whisper_decoder.pte';
-export const WHISPER_TINY_MULTILINGUAL_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-whisper-tiny/resolve/v0.4.0/tokenizer.json';
+const MOONSHINE_TINY_MODEL_NAME = 'moonshine-tiny';
+const MOONSHINE_TINY_ENCODER = `${BASE_URL_PREFIX}-${MOONSHINE_TINY_MODEL_NAME}/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_encoder.pte`;
+const MOONSHINE_TINY_DECODER = `${BASE_URL_PREFIX}-${MOONSHINE_TINY_MODEL_NAME}/${VERSION_TAG}/xnnpack/moonshine_tiny_xnnpack_decoder.pte`;
+const MOONSHINE_TOKENIZER = `${BASE_URL_PREFIX}-${MOONSHINE_TINY_MODEL_NAME}/${VERSION_TAG}/moonshine_tiny_tokenizer.json`;
+
+export const MOONSHINE_TINY = {
+  encoder: MOONSHINE_TINY_ENCODER,
+  decoder: MOONSHINE_TINY_DECODER,
+  tokenizer: MOONSHINE_TOKENIZER,
+};
+
+const WHISPER_TINY_MODEL_NAME = 'whisper-tiny.en';
+const WHISPER_TINY_ENCODER = `${BASE_URL_PREFIX}-${WHISPER_TINY_MODEL_NAME}/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_encoder.pte`;
+const WHISPER_TINY_DECODER = `${BASE_URL_PREFIX}-${WHISPER_TINY_MODEL_NAME}/${VERSION_TAG}/xnnpack/whisper_tiny_en_xnnpack_decoder.pte`;
+const WHISPER_TOKENIZER = `${BASE_URL_PREFIX}-${WHISPER_TINY_MODEL_NAME}/${VERSION_TAG}/whisper_tokenizer.json`;
+
+export const WHISPER_TINY = {
+  encoder: WHISPER_TINY_ENCODER,
+  decoder: WHISPER_TINY_DECODER,
+  tokenizer: WHISPER_TOKENIZER,
+};
+
+const WHISPER_TINY_MULTILINGUAL_MODEL_NAME = 'whisper-tiny';
+const WHISPER_TINY_MULTILINGUAL_ENCODER = `${BASE_URL_PREFIX}-${WHISPER_TINY_MULTILINGUAL_MODEL_NAME}/${VERSION_TAG}/xnnpack/xnnpack_whisper_encoder.pte`;
+const WHISPER_TINY_MULTILINGUAL_DECODER = `${BASE_URL_PREFIX}-${WHISPER_TINY_MULTILINGUAL_MODEL_NAME}/${VERSION_TAG}/xnnpack/xnnpack_whisper_decoder.pte`;
+const WHISPER_TINY_MULTILINGUAL_TOKENIZER = `${BASE_URL_PREFIX}-${WHISPER_TINY_MULTILINGUAL_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+
+export const WHISPER_TINY_MULTILINGUAL = {
+  encoder: WHISPER_TINY_MULTILINGUAL_ENCODER,
+  decoder: WHISPER_TINY_MULTILINGUAL_DECODER,
+  tokenizer: WHISPER_TINY_MULTILINGUAL_TOKENIZER,
+};
 
 // OCR
-export const DETECTOR_CRAFT_1280 =
-  'https://huggingface.co/software-mansion/react-native-executorch-detector-craft/resolve/v0.4.0/xnnpack/xnnpack_craft_1280.pte';
-export const DETECTOR_CRAFT_800 =
-  'https://huggingface.co/software-mansion/react-native-executorch-detector-craft/resolve/v0.4.0/xnnpack/xnnpack_craft_800.pte';
-export const DETECTOR_CRAFT_320 =
-  'https://huggingface.co/software-mansion/react-native-executorch-detector-craft/resolve/v0.4.0/xnnpack/xnnpack_craft_320.pte';
+const DETECTOR_CRAFT_MODEL_NAME = 'detector-craft';
+export const DETECTOR_CRAFT_1280 = `${BASE_URL_PREFIX}-${DETECTOR_CRAFT_MODEL_NAME}/${VERSION_TAG}/xnnpack/xnnpack_craft_1280.pte`;
+export const DETECTOR_CRAFT_800 = `${BASE_URL_PREFIX}-${DETECTOR_CRAFT_MODEL_NAME}/${VERSION_TAG}/xnnpack/xnnpack_craft_800.pte`;
+export const DETECTOR_CRAFT_320 = `${BASE_URL_PREFIX}-${DETECTOR_CRAFT_MODEL_NAME}/${VERSION_TAG}/xnnpack/xnnpack_craft_320.pte`;
 
 // Image segmentation
-export const DEEPLAB_V3_RESNET50 =
-  'https://huggingface.co/software-mansion/react-native-executorch-deeplab-v3/resolve/v0.4.0/xnnpack/deeplabV3_xnnpack_fp32.pte';
+const DEEPLAB_V3_MODEL_NAME = 'deeplab-v3';
+export const DEEPLAB_V3_RESNET50 = `${BASE_URL_PREFIX}-${DEEPLAB_V3_MODEL_NAME}/${VERSION_TAG}/xnnpack/deeplabV3_xnnpack_fp32.pte`;
 
 // Image Embeddings
-export const CLIP_VIT_BASE_PATCH_32_IMAGE_ENCODER_MODEL =
-  'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32-image-encoder/resolve/v0.5.0/clip-vit-base-patch32-image-encoder-float32.pte';
+const CLIP_VIT_BASE_PATCH_32_IMAGE_ENCODER_MODEL_NAME =
+  'clip-vit-base-patch32-image-encoder';
+export const CLIP_VIT_BASE_PATCH_32_IMAGE_ENCODER_MODEL = `${BASE_URL_PREFIX}-${CLIP_VIT_BASE_PATCH_32_IMAGE_ENCODER_MODEL_NAME}/${NEXT_DEV_VERSION_TAG}/clip-vit-base-patch32-image-encoder-float32.pte`;
 
 // Text Embeddings
-export const ALL_MINILM_L6_V2 =
-  'https://huggingface.co/software-mansion/react-native-executorch-all-MiniLM-L6-v2/resolve/v0.4.0/all-MiniLM-L6-v2_xnnpack.pte';
-export const ALL_MINILM_L6_V2_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-all-MiniLM-L6-v2/resolve/v0.4.0/tokenizer.json';
-
-export const ALL_MPNET_BASE_V2 =
-  'https://huggingface.co/software-mansion/react-native-executorch-all-mpnet-base-v2/resolve/v0.4.0/all-mpnet-base-v2_xnnpack.pte';
-export const ALL_MPNET_BASE_V2_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-all-mpnet-base-v2/resolve/v0.4.0/tokenizer.json';
-
-export const MULTI_QA_MINILM_L6_COS_V1 =
-  'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-MiniLM-L6-cos-v1/resolve/v0.4.0/multi-qa-MiniLM-L6-cos-v1_xnnpack.pte';
-export const MULTI_QA_MINILM_L6_COS_V1_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-MiniLM-L6-cos-v1/resolve/v0.4.0/tokenizer.json';
-
-export const MULTI_QA_MPNET_BASE_DOT_V1 =
-  'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-mpnet-base-dot-v1/resolve/v0.4.0/multi-qa-mpnet-base-dot-v1_xnnpack.pte';
-export const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-multi-qa-mpnet-base-dot-v1/resolve/v0.4.0/tokenizer.json';
-
-export const CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_MODEL =
-  'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32-text-encoder/resolve/v0.5.0/clip-vit-base-patch32-text-encoder-float32.pte';
-export const CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_TOKENIZER =
-  'https://huggingface.co/software-mansion/react-native-executorch-clip-vit-base-patch32-text-encoder/resolve/v0.5.0/tokenizer.json';
+const ALL_MINILM_L6_V2_MODEL_NAME = 'all-MiniLM-L6-v2';
+const ALL_MINILM_L6_V2_MODEL = `${BASE_URL_PREFIX}-${ALL_MINILM_L6_V2_MODEL_NAME}/${VERSION_TAG}/all-MiniLM-L6-v2_xnnpack.pte`;
+const ALL_MINILM_L6_V2_TOKENIZER = `${BASE_URL_PREFIX}-${ALL_MINILM_L6_V2_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+
+export const ALL_MINILM_L6_V2 = {
+  modelSource: ALL_MINILM_L6_V2_MODEL,
+  tokenizerSource: ALL_MINILM_L6_V2_TOKENIZER,
+  meanPooling: false,
+};
+
+const ALL_MPNET_BASE_V2_MODEL_NAME = 'all-mpnet-base-v2';
+const ALL_MPNET_BASE_V2_MODEL = `${BASE_URL_PREFIX}-${ALL_MPNET_BASE_V2_MODEL_NAME}/${VERSION_TAG}/all-mpnet-base-v2_xnnpack.pte`;
+const ALL_MPNET_BASE_V2_TOKENIZER = `${BASE_URL_PREFIX}-${ALL_MPNET_BASE_V2_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+
+export const ALL_MPNET_BASE_V2 = {
+  modelSource: ALL_MPNET_BASE_V2_MODEL,
+  tokenizerSource: ALL_MPNET_BASE_V2_TOKENIZER,
+  meanPooling: false,
+};
+
+const MULTI_QA_MINILM_L6_COS_V1_MODEL_NAME = 'multi-qa-MiniLM-L6-cos-v1';
+const MULTI_QA_MINILM_L6_COS_V1_MODEL = `${BASE_URL_PREFIX}-${MULTI_QA_MINILM_L6_COS_V1_MODEL_NAME}/${VERSION_TAG}/multi-qa-MiniLM-L6-cos-v1_xnnpack.pte`;
+const MULTI_QA_MINILM_L6_COS_V1_TOKENIZER = `${BASE_URL_PREFIX}-${MULTI_QA_MINILM_L6_COS_V1_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+
+export const MULTI_QA_MINILM_L6_COS_V1 = {
+  modelSource: MULTI_QA_MINILM_L6_COS_V1_MODEL,
+  tokenizerSource: MULTI_QA_MINILM_L6_COS_V1_TOKENIZER,
+  meanPooling: false,
+};
+
+const MULTI_QA_MPNET_BASE_DOT_V1_MODEL_NAME = 'multi-qa-mpnet-base-dot-v1';
+const MULTI_QA_MPNET_BASE_DOT_V1_MODEL = `${BASE_URL_PREFIX}-${MULTI_QA_MPNET_BASE_DOT_V1_MODEL_NAME}/${VERSION_TAG}/multi-qa-mpnet-base-dot-v1_xnnpack.pte`;
+const MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER = `${BASE_URL_PREFIX}-${MULTI_QA_MPNET_BASE_DOT_V1_MODEL_NAME}/${VERSION_TAG}/tokenizer.json`;
+
+export const MULTI_QA_MPNET_BASE_DOT_V1 = {
+  modelSource: MULTI_QA_MPNET_BASE_DOT_V1_MODEL,
+  tokenizerSource: MULTI_QA_MPNET_BASE_DOT_V1_TOKENIZER,
+  meanPooling: false,
+};
+
+const CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_MODEL_NAME =
+  'clip-vit-base-patch32-text-encoder';
+const CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_MODEL = `${BASE_URL_PREFIX}-${CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_MODEL_NAME}/${NEXT_DEV_VERSION_TAG}/clip-vit-base-patch32-text-encoder-float32.pte`;
+const CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_TOKENIZER = `${BASE_URL_PREFIX}-${CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_MODEL_NAME}/${NEXT_DEV_VERSION_TAG}/tokenizer.json`;
 
 export const CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER = {
   modelSource: CLIP_VIT_BASE_PATCH_32_TEXT_ENCODER_MODEL,
diff --git a/packages/react-native-executorch/src/constants/sttDefaults.ts b/packages/react-native-executorch/src/constants/sttDefaults.ts
index 647c9a1468..48613992a4 100644
--- a/packages/react-native-executorch/src/constants/sttDefaults.ts
+++ b/packages/react-native-executorch/src/constants/sttDefaults.ts
@@ -1,13 +1,7 @@
 import {
-  MOONSHINE_TINY_ENCODER,
-  MOONSHINE_TINY_DECODER,
-  MOONSHINE_TOKENIZER,
-  WHISPER_TINY_ENCODER,
-  WHISPER_TINY_DECODER,
-  WHISPER_TOKENIZER,
-  WHISPER_TINY_MULTILINGUAL_ENCODER,
-  WHISPER_TINY_MULTILINGUAL_DECODER,
-  WHISPER_TINY_MULTILINGUAL_TOKENIZER,
+  MOONSHINE_TINY,
+  WHISPER_TINY,
+  WHISPER_TINY_MULTILINGUAL,
 } from './modelUrls';
 import { AvailableModels, ModelConfig } from '../types/stt';
 
@@ -17,11 +11,11 @@ export const HAMMING_DIST_THRESHOLD = 1;
 
 const whisperTinyModelConfig = {
   sources: {
-    encoder: WHISPER_TINY_ENCODER,
-    decoder: WHISPER_TINY_DECODER,
+    encoder: WHISPER_TINY.encoder,
+    decoder: WHISPER_TINY.decoder,
   },
   tokenizer: {
-    source: WHISPER_TOKENIZER,
+    source: WHISPER_TINY.tokenizer,
     bos: 50257, // FIXME: this is a placeholder and needs to be changed
     eos: 50256, // FIXME: this is a placeholder and needs to be changed
   },
@@ -30,11 +24,11 @@ const whisperTinyModelConfig = {
 
 const moonshineTinyModelConfig = {
   sources: {
-    encoder: MOONSHINE_TINY_ENCODER,
-    decoder: MOONSHINE_TINY_DECODER,
+    encoder: MOONSHINE_TINY.encoder,
+    decoder: MOONSHINE_TINY.decoder,
   },
   tokenizer: {
-    source: MOONSHINE_TOKENIZER,
+    source: MOONSHINE_TINY.tokenizer,
     bos: 1, // FIXME: this is a placeholder and needs to be changed
     eos: 2, // FIXME: this is a placeholder and needs to be changed
   },
@@ -43,11 +37,11 @@ const moonshineTinyModelConfig = {
 
 const whisperTinyMultilingualModelConfig = {
   sources: {
-    encoder: WHISPER_TINY_MULTILINGUAL_ENCODER,
-    decoder: WHISPER_TINY_MULTILINGUAL_DECODER,
+    encoder: WHISPER_TINY_MULTILINGUAL.encoder,
+    decoder: WHISPER_TINY_MULTILINGUAL.decoder,
   },
   tokenizer: {
-    source: WHISPER_TINY_MULTILINGUAL_TOKENIZER,
+    source: WHISPER_TINY_MULTILINGUAL.tokenizer,
     bos: 50258, // FIXME: this is a placeholder and needs to be changed
     eos: 50257, // FIXME: this is a placeholder and needs to be changed
   },