{"success":true,"models":[{"modelId":"zlm-v1-moderation-edge","modelType":"moderation-ensemble","modelVersion":"v1","taskDisplayName":"Text Moderation","taskType":"moderation","onDeviceInference":true,"quantized":true,"quantization":"int8-dynamic","maxTokens":800,"modelDisplayName":"zlm-v1-moderation-edge","parameters":"86M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Text Moderation","Brand Safety","Ad Tech"],"description":"zlm-v1-moderation-edge is a specialized nano model fine-tuned for moderating chat conversations. Similar to OpenAI’s moderation models, it detects unsafe, harmful, and policy-sensitive content and returns structured safety signals that applications can use to block, flag, filter, or route messages. Its compact architecture is optimized for low-latency, high-volume inference, making it well suited as a lightweight safety layer for chat applications, agents, and AI pipelines.","favicon_url":"https://models-favicon.zerogpu.ai/zlm-v1-iab-classify-edge/logo_dark.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/moderations","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.1,"output_per_1m_tokens":0.15,"sample_responses_body":{"input":[{"text":"I am so angry at this person that I want to hurt them. They are worthless and should be scared of what I might do next.","type":"text"}],"model":"zlm-v1-moderation-edge"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"input":[{"text":"I am so angry at this person that I want to hurt them. They are worthless and should be scared of what I might do next.","type":"text"}],"model":"zlm-v1-moderation-edge"}},"displayPriority":24,"createdAt":"2026-08-06T14:57:46+00:00"},{"modelId":"zlm-v1-signal-extract","modelType":"enrichment-seq2seq","modelVersion":"v1","taskDisplayName":"Text Classification","taskType":"signal_extract","onDeviceInference":true,"quantized":true,"quantization":"int8-dynamic","maxTokens":400,"modelDisplayName":"zlm-v1-signal-extract","parameters":"80M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Ad Tech","Chat"],"description":"zlm-v1-signal-extract is a specialized ZeroGPU model for turning unstructured text into structured signals that downstream systems can act on. It extracts topics, keywords, intent, and other contextual attributes in a single inference call, making it well suited for content enrichment, contextual intelligence, ad targeting, agent routing, recommendation systems, and analytics pipelines. The model returns structured output and is designed for fast, high-volume workloads where lightweight signal extraction is more efficient than using a general-purpose LLM.","favicon_url":"https://models-favicon.zerogpu.ai/zlm-v1-iab-classify-edge/logo_dark.png","model_doc_url":"https://docs.zerogpu.ai/","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.04,"output_per_1m_tokens":0.1,"sample_responses_body":{"input":"How can I pay using my credit card?","model":"zlm-v1-signal-extract","categories":["signals"]},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"messages":[{"role":"user","content":"How can I pay using my credit card?"}],"categories":["signals"],"max_keywords":6}},"displayPriority":24,"createdAt":"2026-07-16T04:13:43.125566+00:00"},{"modelId":"whisper-tiny","modelType":null,"modelVersion":"v1","taskDisplayName":"Speech-to-Text","taskType":"transcription","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":null,"modelDisplayName":"whisper-tiny","parameters":"39M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/openai/whisper-tiny","use_cases":["Real-time TTS","Conversational AI","Voice Agents"],"description":"Whisper Tiny is a lightweight automatic speech recognition model from OpenAI designed for fast and efficient speech-to-text transcription. With just 39M parameters, it is the smallest model in the Whisper family and is well suited for high-volume, latency-sensitive, and cost-sensitive audio workloads.\n\nThe multilingual version supports speech recognition across many languages and can also perform speech translation into English. It is ideal for applications such as voice agents, meeting transcription, subtitles, podcasts, audio indexing, and other workloads where speed and efficiency are prioritized over the accuracy of larger speech recognition models.","favicon_url":"https://models-favicon.zerogpu.ai/gpt-oss-120b/gpt-oss-120b.png","model_doc_url":"https://docs.zerogpu.ai","privacy_service":"https://huggingface.co/openai/whisper-tiny","audio_per_minute":0.02,"input_per_1m_tokens":0,"output_per_1m_tokens":0,"sample_responses_body":{"language":"en","temperature":0,"response_format":"verbose_json"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":null},"displayPriority":24,"createdAt":"2026-09-04T14:37:28.487971+00:00"},{"modelId":"zlm-v2-iab-classify-edge-enriched","modelType":"iab-enrichment-classify","modelVersion":"v2","taskDisplayName":"Text Classification","taskType":"iab_classify","onDeviceInference":true,"quantized":false,"quantization":null,"maxTokens":800,"modelDisplayName":"zlm-v2-iab-classify-edge-enriched","parameters":"170M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Ad Tech"],"description":"zlm-v2-iab-classify-edge-enriched is a multilingual content classification model built for contextual intelligence, advertising, and agent workflows. It analyzes text across 50+ languages and returns IAB content categories across the 1.0 and 2.2 taxonomies, along with topics, keywords, audience and interest segments, user intent, and confidence scores. It is optimized for low-latency edge inference and helps ad systems, agents, publishers, and recommendation platforms power contextual targeting, content categorization, brand-safety workflows, and real-time signal enrichment.","favicon_url":"https://models-favicon.zerogpu.ai/zlm-v1-iab-classify-edge-enriched/logo_dark.png","model_doc_url":"https://docs.zerogpu.ai/","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.05,"output_per_1m_tokens":0.15,"sample_responses_body":{"input":"Technology has quietly reshaped the rhythm of everyday life, weaving itself into routines so seamlessly that it often goes unnoticed. From the moment a smartphone alarm wakes someone in the morning to the final glance at a glowing screen before sleep, digital systems guide communication, navigation, work, and entertainment. This transformation did not happen overnight. It emerged through decades of incremental innovation, each new tool building upon the last, until the modern world became deeply interconnected.One of the most significant changes has been the speed at which information travels.","model":"zlm-v2-iab-classify-edge-enriched"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"model":"zlm-v2-iab-classify-edge-enriched","messages":[{"role":"user","content":"What are the basics of exercise and physical fitness?Exercise is anything that gets your body moving. Regular exercise is one of the best things you can do for your health. It has many benefits, including improving your overall health and fitness, and reducing your risk for many chronic (long-term) diseases.Every physical fitness routine is built on a few simple ideas. These include:Make exercise a habit, as your body adapts to the type of activity you do most often. Regular practice will help you improve.Build up your activity level slowly to help you continue to get stronger, faster, or more flexible without pushing too hard all at once.Challenge yourself by lifting slightly heavier weights, adding a few more minutes to your walk, or increasing your pace."}]}},"displayPriority":23,"createdAt":"2026-07-16T04:12:25.299529+00:00"},{"modelId":"chatterbox-nano","modelType":"text-to-speech","modelVersion":"1","taskDisplayName":"Text-to-Speech","taskType":"speech","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":null,"modelDisplayName":"Chatterbox Nano","parameters":"110M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/ResembleAI/chatterbox-nano","use_cases":["Voice Agents","Real-time TTS","Conversational AI"],"description":"Chatterbox Nano is a lightweight text-to-speech model from Resemble AI designed for fast, high-quality speech generation. Its compact architecture makes it well suited for real-time and high-volume workloads, including voice agents, conversational AI, gaming, accessibility, audiobooks, and content creation.\n\nThe model supports expressive speech generation and voice cloning while maintaining low latency and efficient inference. Chatterbox Nano is a strong choice for applications that need natural-sounding audio at significantly lower compute cost than larger TTS models.","favicon_url":"https://cdn-avatars.huggingface.co/v1/production/uploads/6000a0456a2a91af974298cf/qmxSbcVBCwUtVez918IkQ.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/models/chatterbox-nano","privacy_service":"https://huggingface.co/ResembleAI/chatterbox-nano","audio_per_minute":0.01,"input_per_1m_tokens":0,"output_per_1m_tokens":0,"sample_responses_body":{"seed":42,"input":"[dramatic] Oh, another AI model that needs eight GPUs just to say hello? [sarcastic] How adorable. [shush] Do not tell anyone... [happy] I run on the edge.","top_k":1000,"top_p":0.95,"voice":"default","temperature":0.8,"response_format":"mp3","repetition_penalty":1.2},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":null},"displayPriority":23,"createdAt":"2026-09-11T15:16:13.911753+00:00"},{"modelId":"deepseek-v4.1-flash","modelType":"text-generation","modelVersion":"v4.1","taskDisplayName":"Text Generation","taskType":"deepseek_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":1048576,"modelDisplayName":"deepseek-v4.1-flash","parameters":null,"modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/LICENSE","use_cases":["Chat"],"description":"DeepSeek V4.1 Flash is an open-weight sparse mixture-of-experts model and the first built on DeepSeek's Causal Encoder-Decoder (CED) architecture, activating 8B parameters on input and 16B on output. It keeps the one-million-token context window of the V4 Flash line, which suits large codebases, long documents, extended conversations and multi-step agent tasks, and supports function calling alongside both fast non-thinking replies and higher-effort reasoning.","favicon_url":"https://models-favicon.zerogpu.ai/deepseek-ai/deepseek-color.png","model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.3,"output_per_1m_tokens":1.2,"sample_responses_body":{"input":"Hello!","model":"deepseek-v4.1-flash"},"cached_input_per_1m_tokens":0.003,"sample_chat_completions_body":{"model":"deepseek-v4.1-flash","messages":[{"role":"system","content":"You are a support triage analyst. Be concise."},{"role":"user","content":"Classify this ticket by urgency (low/medium/high) and category, then summarize it in one line: 'Our checkout page has been throwing intermittent 502 errors since this morning's deploy. Roughly 1 in 5 customers can't complete payment, and abandoned carts are spiking. Rolling back didn't help."}],"max_tokens":800}},"displayPriority":22,"createdAt":"2026-09-16T13:54:17.230846+00:00"},{"modelId":"glm-5.3-flash","modelType":"text-generation","modelVersion":"v5.3","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":1048576,"modelDisplayName":"glm-5.3-flash","parameters":null,"modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/zai-org/GLM-5.3-Flash/blob/main/LICENSE","use_cases":[],"description":"GLM-5.3-Flash is Z.ai's efficient open-weight model for coding and long-horizon agent tasks. Its hybrid sparse and linear attention keeps long-context behaviour accurate across a one-million-token window while reducing compute, and it supports function calling and adjustable reasoning effort.","favicon_url":"https://cdn-avatars.huggingface.co/v1/production/uploads/62dc173789b4cf157d36ebee/i_pxzM2ZDo3Ub-BEgIkE9.png","model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.1,"output_per_1m_tokens":0.35,"sample_responses_body":{"input":"Hello!","model":"glm-5.3-flash"},"cached_input_per_1m_tokens":0.015,"sample_chat_completions_body":{"model":"glm-5.3-flash","messages":[{"role":"user","content":"Hello!"}]}},"displayPriority":21,"createdAt":"2026-09-18T19:18:26.945544+00:00"},{"modelId":"gliner2.5-base-v1","modelType":"gliner2","modelVersion":"2.5","taskDisplayName":"Text Classification","taskType":"gliner","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":null,"modelDisplayName":"GLiNER2.5 Base v1","parameters":"194M","modelUsecases":{"ner":{"url":null,"description":"Named entity recognition with descriptions and spans","usecase_display_name":"Entity Extraction","sample_responses_body":{"input":"The application is built with Python 3.11 and uses PostgreSQL 15 for storage. It runs on Kubernetes with Docker containers and communicates via gRPC.","model":"gliner25-base-v1","metadata":{"labels":["programming language","database","technology","protocol"],"usecase":"ner","threshold":0.3}},"sample_chat_completions_body":{"model":"gliner25-base-v1","messages":[{"role":"user","content":"The application is built with Python 3.11 and uses PostgreSQL 15 for storage. It runs on Kubernetes with Docker containers and communicates via gRPC."}],"metadata":{"labels":["programming language","database","technology","protocol"],"usecase":"ner","threshold":0.3}}},"json":{"url":null,"description":"Parse complex JSON structures from text","usecase_display_name":"Structured Data Extraction","sample_responses_body":{"input":"Best regards, John Smith, Senior Software Engineer at Acme Corp. Phone: (555) 123-4567, Email: john.smith@acme.com, Office: 123 Main Street, Suite 400, San Francisco, CA 94105","model":"gliner25-base-v1","metadata":{"schema":{"contact":["name::str::Full name","title::str::Job title","company::str::Company name","phone::str::Phone number","email::str::Email address","address::str::Office address"]},"usecase":"json"}},"sample_chat_completions_body":{"model":"gliner25-base-v1","messages":[{"role":"user","content":"Best regards, John Smith, Senior Software Engineer at Acme Corp. Phone: (555) 123-4567, Email: john.smith@acme.com, Office: 123 Main Street, Suite 400, San Francisco, CA 94105"}],"metadata":{"schema":{"contact":["name::str::Full name","title::str::Job title","company::str::Company name","phone::str::Phone number","email::str::Email address","address::str::Office address"]},"usecase":"json"}}},"classification":{"url":null,"description":"Single and multi-label classification with confidence scores","usecase_display_name":"Text Classification","sample_responses_body":{"input":"I absolutely love this product! The quality is outstanding and the customer service was incredibly helpful.","model":"gliner25-base-v1","metadata":{"schema":{"sentiment":["positive","negative","neutral"]},"usecase":"classification"}},"sample_chat_completions_body":{"model":"gliner25-base-v1","messages":[{"role":"user","content":"I absolutely love this product! The quality is outstanding and the customer service was incredibly helpful."}],"metadata":{"schema":{"sentiment":["positive","negative","neutral"]},"usecase":"classification"}}}},"pricing":{"currency":"USD","terms_url":"https://github.com/fastino-ai/GLiNER2/blob/main/LICENSE","use_cases":["PII","Data Extraction"],"description":"GLiNER2.5 Base is an English schema-based information extraction model built on DeBERTa-v3-base. It can handle entity extraction, text classification, structured record extraction, relation extraction, and span-level attributes with a single model.\n\nUseful for GTM workflows like parsing inbound leads, extracting company and contact details, classifying sales conversations, identifying relationships between people and companies, structuring CRM notes, and enriching customer or account data.","favicon_url":"https://models-favicon.zerogpu.ai/gliner2-base-v1/fastino.png","model_doc_url":"https://huggingface.co/fastino/gliner2.5-base-v1","privacy_service":"https://github.com/fastino-ai/GLiNER2/blob/main/LICENSE","audio_per_minute":null,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":null,"cached_input_per_1m_tokens":0,"sample_chat_completions_body":null},"displayPriority":21,"createdAt":"2026-09-22T22:40:13.760384+00:00"},{"modelId":"gpt-4.1-mini","modelType":"text-generation","modelVersion":"v4.1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":1047576,"modelDisplayName":"gpt-4.1-mini","parameters":null,"modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://openai.com/policies/services-agreement/","use_cases":["Coding","Chat","Function Calling","Reasoning","RAG","Summarization","Translation"],"description":"GPT-4.1 mini is OpenAI's fast, cost-efficient GPT-4.1 model. It excels at instruction following and tool calling, supports function calling and structured outputs, and works across a one-million-token context window with low latency.","favicon_url":"https://models-favicon.zerogpu.ai/gpt-oss-120b/gpt-oss-120b.png","model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.4,"output_per_1m_tokens":1.6,"sample_responses_body":{"input":"Hello!","model":"gpt-4.1-mini"},"cached_input_per_1m_tokens":0.1,"sample_chat_completions_body":{"model":"gpt-4.1-mini","messages":[{"role":"user","content":"Hello!"}]}},"displayPriority":20,"createdAt":"2026-09-18T19:18:00.863098+00:00"},{"modelId":"gpt-6-luna","modelType":"text-generation","modelVersion":"v1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":272000,"modelDisplayName":"gpt-6-luna","parameters":null,"modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://openai.com/policies/services-agreement/","use_cases":["Coding","Chat","Function Calling","Reasoning","RAG","Summarization","Translation"],"description":"GPT-6 Luna is the most efficient model of OpenAI's GPT-6 family, built for high-volume, cost-sensitive workloads. It reads text and images, supports adjustable reasoning effort, function calling and structured outputs, and is served through OpenAI's own API.","favicon_url":"https://models-favicon.zerogpu.ai/gpt-oss-120b/gpt-oss-120b.png","model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.1,"output_per_1m_tokens":0.5,"sample_responses_body":{"input":"Hello!","model":"gpt-6-luna"},"cached_input_per_1m_tokens":0.01,"sample_chat_completions_body":{"model":"gpt-6-luna","messages":[{"role":"user","content":"Hello!"}]}},"displayPriority":20,"createdAt":"2026-10-21T17:35:47.858933+00:00"},{"modelId":"gliner2.5-multi-v1","modelType":"gliner2","modelVersion":"2.5","taskDisplayName":"Text Classification","taskType":"gliner","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":null,"modelDisplayName":"GLiNER2.5 Multi v1","parameters":"287M","modelUsecases":{"ner":{"url":null,"description":"Named entity recognition with descriptions and spans","usecase_display_name":"Entity Extraction","sample_responses_body":{"input":"The application is built with Python 3.11 and uses PostgreSQL 15 for storage. It runs on Kubernetes with Docker containers and communicates via gRPC.","model":"gliner25-multi-v1","metadata":{"labels":["programming language","database","technology","protocol"],"usecase":"ner","threshold":0.3}},"sample_chat_completions_body":{"model":"gliner25-multi-v1","messages":[{"role":"user","content":"The application is built with Python 3.11 and uses PostgreSQL 15 for storage. It runs on Kubernetes with Docker containers and communicates via gRPC."}],"metadata":{"labels":["programming language","database","technology","protocol"],"usecase":"ner","threshold":0.3}}},"json":{"url":null,"description":"Parse complex JSON structures from text","usecase_display_name":"Structured Data Extraction","sample_responses_body":{"input":"Best regards, John Smith, Senior Software Engineer at Acme Corp. Phone: (555) 123-4567, Email: john.smith@acme.com, Office: 123 Main Street, Suite 400, San Francisco, CA 94105","model":"gliner25-multi-v1","metadata":{"schema":{"contact":["name::str::Full name","title::str::Job title","company::str::Company name","phone::str::Phone number","email::str::Email address","address::str::Office address"]},"usecase":"json"}},"sample_chat_completions_body":{"model":"gliner25-multi-v1","messages":[{"role":"user","content":"Best regards, John Smith, Senior Software Engineer at Acme Corp. Phone: (555) 123-4567, Email: john.smith@acme.com, Office: 123 Main Street, Suite 400, San Francisco, CA 94105"}],"metadata":{"schema":{"contact":["name::str::Full name","title::str::Job title","company::str::Company name","phone::str::Phone number","email::str::Email address","address::str::Office address"]},"usecase":"json"}}},"classification":{"url":null,"description":"Single and multi-label classification with confidence scores","usecase_display_name":"Text Classification","sample_responses_body":{"input":"I absolutely love this product! The quality is outstanding and the customer service was incredibly helpful.","model":"gliner25-multi-v1","metadata":{"schema":{"sentiment":["positive","negative","neutral"]},"usecase":"classification"}},"sample_chat_completions_body":{"model":"gliner25-multi-v1","messages":[{"role":"user","content":"I absolutely love this product! The quality is outstanding and the customer service was incredibly helpful."}],"metadata":{"schema":{"sentiment":["positive","negative","neutral"]},"usecase":"classification"}}}},"pricing":{"currency":"USD","terms_url":"https://github.com/fastino-ai/GLiNER2/blob/main/LICENSE","use_cases":["PII","Data Extraction"],"description":"GLiNER2.5 Multi is a 287M-parameter multilingual information extraction model built on mDeBERTa-v3-base. It supports zero-shot named entity recognition, text classification, and structured JSON extraction across multiple languages, making it well suited for extracting people, organizations, locations, attributes, and other custom entities without task-specific retraining.","favicon_url":"https://models-favicon.zerogpu.ai/gliner2-base-v1/fastino.png","model_doc_url":"https://huggingface.co/fastino/gliner2.5-multi-v1","privacy_service":"https://github.com/fastino-ai/GLiNER2/blob/main/LICENSE","audio_per_minute":null,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":null,"cached_input_per_1m_tokens":0,"sample_chat_completions_body":null},"displayPriority":20,"createdAt":"2026-09-22T22:45:51.085892+00:00"},{"modelId":"gpt-5.6-luna","modelType":"text-generation","modelVersion":"v1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":272000,"modelDisplayName":"gpt-5.6-luna","parameters":null,"modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://openai.com/policies/services-agreement/","use_cases":["Coding","Chat","Function Calling","Reasoning","RAG","Summarization","Translation"],"description":"GPT-5.6 Luna is the cost-optimized model of OpenAI's GPT-5.6 family, designed for cost-sensitive, high-volume workloads. It supports adjustable reasoning effort, function calling and structured outputs.","favicon_url":"https://models-favicon.zerogpu.ai/gpt-oss-120b/gpt-oss-120b.png","model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.2,"output_per_1m_tokens":1.2,"sample_responses_body":{"input":"Hello!","model":"gpt-5.6-luna"},"cached_input_per_1m_tokens":0.2,"sample_chat_completions_body":{"model":"gpt-5.6-luna","messages":[{"role":"user","content":"Hello!"}]}},"displayPriority":20,"createdAt":"2026-09-18T19:18:00.863098+00:00"},{"modelId":"gpt-5.4-nano","modelType":"text-generation","modelVersion":"v1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":400000,"modelDisplayName":"gpt-5.4-nano","parameters":null,"modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://openai.com/policies/services-agreement/","use_cases":["Coding","Chat","Function Calling","Reasoning","RAG","Summarization","Translation"],"description":"GPT-5.4 nano is the most cost-efficient model in OpenAI's GPT-5.4 family, built for high-volume and latency-sensitive workloads such as classification, extraction, routing and sub-agent tasks. It supports function calling and structured outputs with a 400,000-token context window.","favicon_url":"https://models-favicon.zerogpu.ai/gpt-oss-120b/gpt-oss-120b.png","model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.2,"output_per_1m_tokens":1.25,"sample_responses_body":{"input":"Hello!","model":"gpt-5.4-nano"},"cached_input_per_1m_tokens":0.02,"sample_chat_completions_body":{"model":"gpt-5.4-nano","messages":[{"role":"user","content":"Hello!"}]}},"displayPriority":20,"createdAt":"2026-09-18T19:18:00.863098+00:00"},{"modelId":"gpt-oss-120b","modelType":"text-generation","modelVersion":"1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":131072,"modelDisplayName":"gpt-oss-120b","parameters":"120B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://github.com/openai/gpt-oss/blob/main/LICENSE","use_cases":["Coding","Chat","Function Calling","Reasoning","RAG","Summarization","Translation"],"description":"gpt-oss-120b is OpenAI’s open-weight reasoning model built for advanced text generation, coding, research, and agentic workflows. It supports function calling, structured outputs, streaming, and multilingual use, making it a strong fit for applications that need deeper reasoning and reliable multi-step execution. On ZeroGPU, gpt-oss-120b runs across the hybrid inference cloud, using optimized cloud and edge capacity to deliver strong performance while balancing speed and cost for production workloads.","favicon_url":"https://models-favicon.zerogpu.ai/gpt-oss-120b/gpt-oss-120b.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/models/gpt-oss-120b","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.15,"output_per_1m_tokens":0.6,"sample_responses_body":{"input":"Name my WiFi network something tha will make my neighbors laugh. Give me your top 3 with a one-line reason each.","model":"gpt-oss-120b","max_tokens":800,"instructions":"You are a witty naming consultant. Be brief."},"cached_input_per_1m_tokens":0.03,"sample_chat_completions_body":{"model":"gpt-oss-120b","messages":[{"role":"system","content":"You are a witty naming consultant. Be brief."},{"role":"user","content":"Name my WiFi network something that will make my neighbors laugh. Give me your top 3 with a one-line reason each."}],"max_tokens":800}},"displayPriority":18,"createdAt":"2026-07-17T08:01:26.891155+00:00"},{"modelId":"qwen3-30b-a3b-fp8","modelType":"text-generation","modelVersion":"1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":true,"quantization":"fp8","maxTokens":32768,"modelDisplayName":"qwen3-30b-a3b-fp8","parameters":"30B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/Qwen/Qwen3-30B-A3B-FP8/blob/main/LICENSE","use_cases":["Coding","Chat","Function Calling","Reasoning","RAG","Summarization","Translation"],"description":"qwen3-30b-a3b-fp8 is Qwen’s open-weight mixture-of-experts model designed for efficient reasoning, coding, multilingual text generation, and agentic workflows. It activates only 3B of its 30B parameters per token, delivering strong model quality with lower latency and inference cost. It supports function calling, streaming, and multilingual workloads, making it well suited for agents, code assistance, automation, structured data tasks, and high-volume production applications. On ZeroGPU, it runs across the hybrid inference cloud to optimize speed and cost for production workloads.","favicon_url":"https://models-favicon.zerogpu.ai/qwen3-30b-a3b-fp8/qwen3-30b-a3b-fp8.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/models/qwen3-30b-a3b-fp8","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.1,"output_per_1m_tokens":0.45,"sample_responses_body":null,"cached_input_per_1m_tokens":0.05,"sample_chat_completions_body":{"model":"qwen3-30b-a3b-fp8","messages":[{"role":"system","content":"You are a pragmatic senior engineer. Be brief."},{"role":"user","content":"My API is getting rate-limited by a third party. Give me 3\n  battle-tested strategies to handle it, one line each."}],"max_tokens":800}},"displayPriority":17,"createdAt":"2026-07-17T21:55:48.407315+00:00"},{"modelId":"llama-3.2-3b-instruct-fast","modelType":"text-generation","modelVersion":"1","taskDisplayName":"Summarization","taskType":"summary","onDeviceInference":true,"quantized":true,"quantization":"Q4_K_L","maxTokens":131072,"modelDisplayName":"llama-3.2-3b-instruct-fast","parameters":"3B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Summarization","Chat","RAG","Translation","Intent Detection"],"description":"llama-3.2-3b-instruct-fast is Meta's Llama 3.2 3B Instruct model packaged by ZeroGPU for fast, low-cost summarization on edge GPUs. It shares the 128K context and the summarization pipeline of llama-3.1-8b-instruct-fast and runs on ZeroGPU worker devices with GPU acceleration, falling back to the hybrid inference cloud when no device is available.","favicon_url":"https://models-favicon.zerogpu.ai/llama-3.2-3b-instruct/meta.png","model_doc_url":"https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.05,"output_per_1m_tokens":0.3,"sample_responses_body":null,"cached_input_per_1m_tokens":0.025,"sample_chat_completions_body":{"model":"llama-3.2-3b-instruct-fast","messages":[{"role":"user","content":"The global semiconductor industry is undergoing one of its most significant structural shifts in decades, driven by a combination of geopolitical tensions, surging demand from artificial intelligence workloads, and a wave of government-backed industrial policy across the United States, Europe, and Asia. At the center of this transformation is a race to onshore chip manufacturing capacity that had, for thirty years, been quietly concentrated in Taiwan and South Korea. The United States CHIPS and Science Act, signed into law in 2022, allocated over $52 billion in subsidies to incentivize domestic semiconductor fabrication. Since then, companies including TSMC, Intel, Samsung, and Micron have announced or broken ground on new fabs in Arizona, Ohio, Texas, and Idaho. However, construction timelines have slipped, costs have ballooned, and a shortage of skilled workers has prompted some manufacturers to bring in engineers from overseas — a move that has drawn political scrutiny even as it addresses a genuine talent gap. Meanwhile, the explosion of AI model training and inference has fundamentally altered the demand profile for chips. Graphics processing units originally designed for gaming, particularly those made by NVIDIA, have become the primary compute substrate for large language models. NVIDIA's H100 and successor Blackwell-series GPUs now trade at significant premiums on secondary markets, with some hyperscalers reporting lead times of over a year for large cluster orders. This bottleneck has accelerated investment in custom silicon: Google's Tensor Processing Units, Amazon's Trainium and Inferentia chips, and Meta's MTIA accelerators are all attempts to reduce dependence on third-party GPU supply chains. On the geopolitical front, the United States has progressively tightened export controls on advanced chips and chip-making equipment destined for China, with the most recent rules targeting gate-all-around transistor technology and high-bandwidth memory. China, in response, has dramatically increased state investment in its domestic semiconductor ecosystem, with companies like SMIC and Huawei making incremental but closely watched progress at advanced nodes. Analysts remain divided on how quickly China can close the gap, with estimates ranging from five to fifteen years depending on assumptions about technology transfer, equipment access, and the pace of indigenous innovation. The equipment layer — often overlooked in public discourse — may prove to be the most consequential chokepoint. ASML, the Dutch company that holds a near-monopoly on extreme ultraviolet lithography machines, has become a focal point of both trade diplomacy and industrial strategy. Without EUV tools, fabricating chips below 7 nanometers at scale is effectively impossible, giving ASML and its supply chain extraordinary leverage in a world where nations are suddenly treating semiconductor access as a matter of national security."}]}},"displayPriority":16,"createdAt":"2026-10-07T19:35:18.399628+00:00"},{"modelId":"llama-3.1-8b-instruct-fast","modelType":"text-generation","modelVersion":"2","taskDisplayName":"Summarization","taskType":"cloudflare_workers_ai","onDeviceInference":true,"quantized":true,"quantization":"Q4_K_M","maxTokens":131072,"modelDisplayName":"llama-3.1-8b-instruct-fast","parameters":"8B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Summarization","Chat","RAG","Translation","Intent Detection"],"description":"llama-3.1-8b-instruct-fast is Meta’s Llama 3.1 8B Instruct model optimized by ZeroGPU for fast, cost-efficient summarization and text processing at scale. Its 128K context window makes it well suited for long documents, transcripts, articles, email threads, and conversations that need to be processed in a single pass. The model supports multilingual workloads and is a strong fit for summarization pipelines, content processing, and agent workflows where low latency and predictable inference cost matter. On ZeroGPU, it runs across the hybrid inference cloud to optimize speed and cost for high-volume production workloads.","favicon_url":"https://models-favicon.zerogpu.ai/llama-3.2-3b-instruct/meta.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/models/llama-3-1-8b-instruct-fast","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.15,"output_per_1m_tokens":0.28,"sample_responses_body":null,"cached_input_per_1m_tokens":0.025,"sample_chat_completions_body":{"model":"llama-3.1-8b-instruct-fast","messages":[{"role":"user","content":"The global semiconductor industry is undergoing one of its most significant structural shifts in decades, driven by a combination of geopolitical tensions, surging demand from artificial intelligence workloads, and a wave of government-backed industrial policy across the United States, Europe, and Asia. At the center of this transformation is a race to onshore chip manufacturing capacity that had, for thirty years, been quietly concentrated in Taiwan and South Korea. The United States CHIPS and Science Act, signed into law in 2022, allocated over $52 billion in subsidies to incentivize domestic semiconductor fabrication. Since then, companies including TSMC, Intel, Samsung, and Micron have announced or broken ground on new fabs in Arizona, Ohio, Texas, and Idaho. However, construction timelines have slipped, costs have ballooned, and a shortage of skilled workers has prompted some manufacturers to bring in engineers from overseas — a move that has drawn political scrutiny even as it addresses a genuine talent gap. Meanwhile, the explosion of AI model training and inference has fundamentally altered the demand profile for chips. Graphics processing units originally designed for gaming, particularly those made by NVIDIA, have become the primary compute substrate for large language models. NVIDIA's H100 and successor Blackwell-series GPUs now trade at significant premiums on secondary markets, with some hyperscalers reporting lead times of over a year for large cluster orders. This bottleneck has accelerated investment in custom silicon: Google's Tensor Processing Units, Amazon's Trainium and Inferentia chips, and Meta's MTIA accelerators are all attempts to reduce dependence on third-party GPU supply chains. On the geopolitical front, the United States has progressively tightened export controls on advanced chips and chip-making equipment destined for China, with the most recent rules targeting gate-all-around transistor technology and high-bandwidth memory. China, in response, has dramatically increased state investment in its domestic semiconductor ecosystem, with companies like SMIC and Huawei making incremental but closely watched progress at advanced nodes. Analysts remain divided on how quickly China can close the gap, with estimates ranging from five to fifteen years depending on assumptions about technology transfer, equipment access, and the pace of indigenous innovation. The equipment layer — often overlooked in public discourse — may prove to be the most consequential chokepoint. ASML, the Dutch company that holds a near-monopoly on extreme ultraviolet lithography machines, has become a focal point of both trade diplomacy and industrial strategy. Without EUV tools, fabricating chips below 7 nanometers at scale is effectively impossible, giving ASML and its supply chain extraordinary leverage in a world where nations are suddenly treating semiconductor access as a matter of national security."}]}},"displayPriority":16,"createdAt":"2026-05-26T16:34:16.809145+00:00"},{"modelId":"deberta-v3-small","modelType":"nli-deberta-v3","modelVersion":"1","taskDisplayName":"Text Classification","taskType":"classification","onDeviceInference":true,"quantized":true,"quantization":"onnx-int-q8","maxTokens":400,"modelDisplayName":"deberta-v3-small","parameters":"142M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://github.com/microsoft/DeBERTa/blob/master/SECURITY.md","use_cases":["Email Filtering","Intent Detection"],"description":"deberta-v3-small is Microsoft’s lightweight text classification model optimized for fast, low-cost zero-shot classification. Provide text together with your own candidate labels, and the model returns confidence scores for each category without requiring a fixed taxonomy or task-specific training. It is well suited for intent detection, content tagging, routing, filtering, prioritization, and other high-volume classification workflows where speed and cost matter. Its compact architecture makes it a strong alternative to using a general-purpose LLM for straightforward classification tasks.","favicon_url":"https://models-favicon.zerogpu.ai/nli-deberta-v3-small/icons8-microsoft-96.png","model_doc_url":"https://github.com/microsoft/DeBERTa/tree/master/docs","privacy_service":"https://github.com/microsoft/DeBERTa/blob/master/LICENSE","audio_per_minute":0,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":{"input":"Apple is expected to unveil its next-generation M5 chip at WWDC this June, promising a 40% boost in GPU performance and a new dedicated AI core for on-device machine learning tasks.","model":"deberta-v3-small","instructions":"[sports, finance, politics]"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"model":"deberta-v3-small","messages":[{"role":"system","content":"[Sports,Politics,Technology,Finance,Entertainment]"},{"role":"user","content":"Apple is expected to unveil its next-generation M5 chip at WWDC this June, promising a 40% boost in GPU performance and a new dedicated AI core for on-device machine learning tasks."}]}},"displayPriority":15,"createdAt":"2026-03-09T12:09:06.022639+00:00"},{"modelId":"llama-guard-4-12b","modelType":"text-classification","modelVersion":"4","taskDisplayName":"Text Generation","taskType":"moderation","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":163840,"modelDisplayName":"llama-guard-4-12b","parameters":"12B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/meta-llama/Llama-Guard-4-12B","use_cases":["Brand Safety","Text Moderation"],"description":"Llama Guard 4 12B is Meta’s multimodal safety classification model for moderating text, images, and mixed text-image inputs. It can evaluate both incoming prompts and generated responses, returning safe or unsafe classifications along with the relevant policy categories when violations are detected. The model supports multilingual text and multiple images, making it well suited for chat moderation, multimodal AI safety, prompt and response filtering, policy enforcement, and agent guardrails. It is built on a dense 12B architecture derived from Llama 4 Scout and is designed as a dedicated safety layer for production AI applications.","favicon_url":"https://models-favicon.zerogpu.ai/llama-3.2-3b-instruct/meta.png","model_doc_url":"https://huggingface.co/meta-llama/Llama-Guard-4-12B","privacy_service":"https://huggingface.co/meta-llama/Llama-Guard-4-12B","audio_per_minute":null,"input_per_1m_tokens":0.18,"output_per_1m_tokens":0.18,"sample_responses_body":{"input":[{"text":"I am so angry at this person that I want to hurt them. They are worthless and should be scared of what I might do next.","type":"text"}],"model":"llama-guard-4-12b"},"cached_input_per_1m_tokens":null,"sample_chat_completions_body":null},"displayPriority":15,"createdAt":"2026-09-11T15:16:17.041016+00:00"},{"modelId":"zlm-v1-iab-domain-classifier","modelType":"iab-classify","modelVersion":"1","taskDisplayName":"Text Classification","taskType":"iab_classify_domain","onDeviceInference":true,"quantized":true,"quantization":"int8-dynamic","maxTokens":100,"modelDisplayName":"zlm-v1-iab-domain-classifier","parameters":"149M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Ad Tech","Chat"],"description":"zlm-v1-iab-domain-classifier is a low-latency domain classification model built for advertising, contextual intelligence, and large-scale enrichment workflows. It takes a raw domain name as input and returns structured IAB content categories, topics, keywords, and user-intent signals without requiring page content. This makes it well suited for bidstream enrichment, contextual targeting, domain intelligence, publisher classification, and other high-volume adtech workflows where fast, lightweight classification is needed.","favicon_url":"https://models-favicon.zerogpu.ai/zlm-v1-iab-domain-classifier/logo_dark.png","model_doc_url":"https://docs.zerogpu.ai/","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":{"input":"coursera.com","model":"zlm-v1-iab-domain-classifier"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"model":"zlm-v1-iab-domain-classifier","messages":[{"role":"user","content":"coursera.com"}]}},"displayPriority":8,"createdAt":"2026-06-25T14:25:23+00:00"},{"modelId":"zlm-v1-iab-classify-edge","modelType":"iab-classify","modelVersion":"1","taskDisplayName":"Text Classification","taskType":"iab_classify","onDeviceInference":true,"quantized":true,"quantization":"int8-dynamic","maxTokens":400,"modelDisplayName":"zlm-v1-iab-classify-edge","parameters":"90M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://zerogpu.ai/terms","use_cases":["Ad Tech"],"description":"ZeroGPU's IAB classifier maps any text to the industry-standard IAB Content Taxonomy in a single, fast inference call. Each call returns categories across both the 1.0 and 2.2 taxonomies plus matched audience segments, every result scored for confidence. At 90M parameters on ONNX, it's built for the high-volume, low-latency classification that ad tech and content platforms demand, right at the edge. When you need clean category and audience signals on every request, this is the model.","favicon_url":"https://models-favicon.zerogpu.ai/zlm-v1-iab-classify-edge/logo_dark.png","model_doc_url":"https://docs.zerogpu.ai/","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":0,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":{"input":"Technology has quietly reshaped the rhythm of everyday life, weaving itself into routines so seamlessly that it often goes unnoticed. From the moment a smartphone alarm wakes someone in the morning to the final glance at a glowing screen before sleep, digital systems guide communication, navigation, work, and entertainment. This transformation did not happen overnight. It emerged through decades of incremental innovation, each new tool building upon the last, until the modern world became deeply interconnected.One of the most significant changes has been the speed at which information travels.","model":"zlm-v1-iab-classify-edge"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"model":"zlm-v1-iab-classify-edge","messages":[{"role":"user","content":"What are the basics of exercise and physical fitness?Exercise is anything that gets your body moving. Regular exercise is one of the best things you can do for your health. It has many benefits, including improving your overall health and fitness, and reducing your risk for many chronic (long-term) diseases.Every physical fitness routine is built on a few simple ideas. These include:Make exercise a habit, as your body adapts to the type of activity you do most often. Regular practice will help you improve.Build up your activity level slowly to help you continue to get stronger, faster, or more flexible without pushing too hard all at once.Challenge yourself by lifting slightly heavier weights, adding a few more minutes to your walk, or increasing your pace."}]}},"displayPriority":4,"createdAt":"2026-03-04T16:10:05.376044+00:00"},{"modelId":"LFM2.5-1.2B-Thinking","modelType":"lfm2.5","modelVersion":"2.5","taskDisplayName":"Text Generation","taskType":"text_generation","onDeviceInference":false,"quantized":true,"quantization":"Q4_K_M","maxTokens":32768,"modelDisplayName":"LFM2.5-1.2B-Thinking","parameters":"1.2B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://www.liquid.ai/terms-conditions","use_cases":["Intent Detection","Translation","Summarization","Reasoning","Function Calling"],"description":"LFM2.5-1.2B-Thinking is Liquid AI’s compact 1.2B-parameter reasoning model designed for multi-step problem solving, planning, data extraction, and agentic workflows. Its small footprint and 32K context window make it well suited for low-latency inference at the edge, especially for high-volume tasks that need more reasoning than a simple classifier. It is a strong fit for agent planning, routing, tool orchestration, structured extraction, and lightweight reasoning workloads where speed and cost matter. For knowledge-heavy or complex coding tasks, a larger general-purpose model may be a better fit.","favicon_url":"https://models-favicon.zerogpu.ai/LFM2.5-1.2B-Thinking/LFM2.5-1.2B-Thinking_liquid_ai_logo.png","model_doc_url":"https://docs.liquid.ai/deployment/on-device/android/ai-agent-usage-guide#text-models","privacy_service":"https://www.liquid.ai/lfm-license","audio_per_minute":0,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":{"input":"Our whole team got locked out of the dashboard this morning after the payment failed. We have a product demo with a customer in 2 hours and cannot log in. Card on file should be valid.","model":"LFM2.5-1.2B-Thinking","instructions":"You are a support triage assistant. Read the ticket, briefly reason about the signals, then assign one team (Billing, Technical, Account, General) and one priority (Low, Medium, High). Decide once and commit. End with a single line: TEAM | PRIORITY."},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"model":"LFM2.5-1.2B-Thinking","messages":[{"role":"system","content":"You are a support triage assistant. Read the ticket, briefly reason about the signals, then assign one team (Billing, Technical, Account, General) and one priority (Low, Medium, High). Decide once and commit. End with a single line: TEAM | PRIORITY."},{"role":"user","content":"Our whole team got locked out of the dashboard this morning after the payment failed. We have a product demo with a customer in 2 hours and cannot log in. Card on file should be valid."}]}},"displayPriority":3,"createdAt":"2026-03-23T17:52:50+00:00"},{"modelId":"LFM2.5-1.2B-Instruct","modelType":"lfm2.5","modelVersion":"2.5","taskDisplayName":"Text Generation","taskType":"text_generation","onDeviceInference":false,"quantized":true,"quantization":"Q4_K_M","maxTokens":32768,"modelDisplayName":"LFM2.5-1.2B-Instruct","parameters":"1.2B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://www.liquid.ai/terms-conditions","use_cases":["Intent Detection","Translation","Summarization","Chat","Function Calling"],"description":"LFM2.5-1.2B-Instruct is Liquid AI’s compact 1.2B-parameter instruction model built for efficient text generation, conversational AI, and lightweight agent workflows. It supports native tool calling and multilingual use, while its small footprint makes it well suited for low-latency inference on edge devices and resource-constrained environments. It is a strong fit for chat, simple assistants, instruction following, routing, and high-volume agent tasks where speed, efficiency, and cost matter more than deep reasoning.","favicon_url":"https://models-favicon.zerogpu.ai/LFM2.5-1.2B-Instruct/liquid_ai_logo.png","model_doc_url":"https://docs.liquid.ai/deployment/on-device/android/ai-agent-usage-guide#text-models","privacy_service":"https://www.liquid.ai/lfm-license","audio_per_minute":0,"input_per_1m_tokens":0.02,"output_per_1m_tokens":0.05,"sample_responses_body":{"input":"I just completed a 5K run in 28 minutes. Give me a short motivational follow-up message.","model":"LFM2.5-1.2B-Instruct","instructions":"You are a friendly in-app assistant for a fitness app."},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"model":"LFM2.5-1.2B-Instruct","messages":[{"role":"system","content":"You are a friendly in-app assistant for a fitness app."},{"role":"user","content":"I just completed a 5K run in 28 minutes. Give me a short motivational follow-up message."}]}},"displayPriority":2,"createdAt":"2026-03-26T17:18:47.56526+00:00"},{"modelId":"qwen3.6-35b-a3b","modelType":"image-text-to-text","modelVersion":"1","taskDisplayName":"Text Generation","taskType":"text_generation","onDeviceInference":false,"quantized":false,"quantization":"fp8","maxTokens":16384,"modelDisplayName":"Qwen3.6 35B A3B","parameters":"35B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","use_cases":["Chat","Document Retrieval","Data Extraction","Video Analysis","Image Analysis"],"description":"Qwen3.6-35B-A3B is a fast, open-weight multimodal model for applications that need to understand text, images, and video. It’s a strong choice for visual question answering, document and screenshot analysis, video understanding, content extraction, agent workflows, and tool calling. With a 262K context window and structured output support, it works well for processing large amounts of multimodal content while keeping inference efficient with its Mixture-of-Experts architecture.","favicon_url":"https://cdn-avatars.huggingface.co/v1/production/uploads/6215ca5692c0ecfba9186921/hrRM50-6XcdWgg2AKpENG.jpeg","model_doc_url":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","privacy_service":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","audio_per_minute":null,"input_per_1m_tokens":0.2,"output_per_1m_tokens":1.5,"sample_responses_body":{"input":[{"role":"user","content":[{"type":"input_image","detail":"auto","image_url":"https://upload.wikimedia.org/wikipedia/commons/3/3a/Cat03.jpg"},{"text":"What animal is this, and what is it doing?","type":"input_text"}]}],"model":"qwen3.6-35b-a3b","temperature":0.2,"max_output_tokens":512},"cached_input_per_1m_tokens":0.2,"sample_chat_completions_body":{"model":"qwen3.6-35b-a3b","messages":[{"role":"user","content":[{"type":"video_url","video_url":{"url":"https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen3.5/demo/video/N1cdUjctpG8.mp4"}},{"type":"image_url","image_url":{"url":"https://upload.wikimedia.org/wikipedia/commons/3/3a/Cat03.jpg","detail":"auto"}},{"text":"Describe what happens in the video, then say whether the animal in the photo appears in it.","type":"text"}]}],"max_tokens":512,"temperature":0.2}},"displayPriority":0,"createdAt":"2026-09-25T20:13:41.311821+00:00"},{"modelId":"gliner2.5-decide-v1","modelType":"text-classification","modelVersion":"1","taskDisplayName":"Text Classification","taskType":"gliner","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":null,"modelDisplayName":"GLiNER2.5 Decide v1","parameters":"486M","modelUsecases":{"decide":{"url":null,"description":"Dynamic-label decisions across one or more heads: intent, routing, triage, handoff, moderation","usecase_display_name":"Decision / Routing","sample_responses_body":{"input":"The customer says their payment failed and they want someone to call them.","model":"gliner2.5-decide-v1","metadata":{"labels":{"intent":["payment_issue","refund","account_access","general_question"],"handoff":["yes","no"]},"usecase":"decide"}},"sample_chat_completions_body":{"model":"gliner2.5-decide-v1","messages":[{"role":"user","content":"The customer says their payment failed and they want someone to call them."}],"metadata":{"labels":{"intent":["payment_issue","refund","account_access","general_question"],"handoff":["yes","no"]},"usecase":"decide"}}},"classification":{"url":null,"description":"Single- or multi-label text classification against a caller-supplied schema","usecase_display_name":"Text Classification","sample_responses_body":{"input":"I absolutely love this product! The quality is outstanding and the customer service was incredibly helpful.","model":"gliner2.5-decide-v1","metadata":{"schema":{"sentiment":["positive","negative","neutral"]},"usecase":"classification"}},"sample_chat_completions_body":{"model":"gliner2.5-decide-v1","messages":[{"role":"user","content":"I absolutely love this product! The quality is outstanding and the customer service was incredibly helpful."}],"metadata":{"schema":{"sentiment":["positive","negative","neutral"]},"usecase":"classification"}}}},"pricing":{"currency":"USD","terms_url":"https://github.com/fastino-ai/GLiNER2/blob/main/LICENSE","use_cases":["PII","Data Extraction"],"description":"GLiNER2.5 Decide is a lightweight, zero-shot decision and classification model built for operational AI workflows. Dynamically supply labels or decision options at request time to classify intent, route agents and tools, triage support requests, make moderation decisions, detect completion, and determine when human handoff is needed.\nSupports single-label, multi-label, and multiple decision heads in a single call, making it well suited for high-volume routing and decision workloads where speed and structured outputs matter.","favicon_url":"https://models-favicon.zerogpu.ai/gliner2-base-v1/fastino.png","model_doc_url":"https://huggingface.co/fastino/GLiNER2.5-Decide","privacy_service":"https://github.com/fastino-ai/GLiNER2/blob/main/LICENSE","audio_per_minute":null,"input_per_1m_tokens":0.004,"output_per_1m_tokens":0,"sample_responses_body":null,"cached_input_per_1m_tokens":null,"sample_chat_completions_body":null},"displayPriority":0,"createdAt":"2026-09-30T21:10:45.578389+00:00"},{"modelId":"zlm-v1-tool-decisions","modelType":"text-generation","modelVersion":"v1","taskDisplayName":"Text Generation","taskType":"cloudflare_workers_ai","onDeviceInference":false,"quantized":false,"quantization":"bf16","maxTokens":24576,"modelDisplayName":"zlm-v1-tool-decisions","parameters":"0.6B","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/Qwen/Qwen3-0.6B/blob/main/LICENSE","use_cases":[],"description":"zlm-v1-tool-decisions turns user intent into action, combining fast MCP tool selection with high-quality query and argument generation.\n\nBuilt to help agents complete tasks efficiently, it understands the conversation, selects the right tools, and generates the inputs needed to use them. Decision-making and text generation work together in a single inference step, helping agents move from understanding a request to taking action.\n\nWhen no tool is needed, it returns NO_TOOL, avoiding unnecessary calls and keeping workflows efficient. Purpose-built for agents that need to choose quickly and act precisely.","favicon_url":null,"model_doc_url":"https://docs.zerogpu.ai/docs/model-catalog","privacy_service":"https://zerogpu.ai/privacy-policy","audio_per_minute":null,"input_per_1m_tokens":0.05,"output_per_1m_tokens":0.15,"sample_responses_body":{"input":"Will it rain in Austin this week? And what's the latest news on the Longhorns?","model":"zlm-v1-tool-decisions","tools":[{"name":"web_search","type":"function","parameters":{"type":"object","required":["query"],"properties":{"query":{"type":"string","description":"Search query"}}},"description":"Search the web for current information: news, facts, events and prices."},{"name":"get_forecast","type":"function","parameters":{"type":"object","required":["city"],"properties":{"city":{"type":"string","description":"City name"}}},"description":"7-day weather forecast for a city."},{"name":"get_stock_quote","type":"function","parameters":{"type":"object","required":["ticker"],"properties":{"ticker":{"type":"string","description":"Ticker symbol, e.g. AAPL"}}},"description":"Real-time stock quote for an exchange ticker symbol."}],"temperature":0},"cached_input_per_1m_tokens":null,"sample_chat_completions_body":{"model":"zlm-v1-tool-decisions","tools":[{"type":"function","function":{"name":"web_search","parameters":{"type":"object","required":["query"],"properties":{"query":{"type":"string","description":"Search query"}}},"description":"Search the web for current information: news, facts, events and prices."}},{"type":"function","function":{"name":"get_forecast","parameters":{"type":"object","required":["city"],"properties":{"city":{"type":"string","description":"City name"}}},"description":"7-day weather forecast for a city."}},{"type":"function","function":{"name":"get_stock_quote","parameters":{"type":"object","required":["ticker"],"properties":{"ticker":{"type":"string","description":"Ticker symbol, e.g. AAPL"}}},"description":"Real-time stock quote for an exchange ticker symbol."}}],"messages":[{"role":"user","content":"Will it rain in Austin this week? And what's the latest news on the Longhorns?"}],"temperature":0}},"displayPriority":0,"createdAt":"2026-10-26T17:33:56.6793+00:00"},{"modelId":"all-minilm-l6-v2","modelType":"BERT","modelVersion":"v1","taskDisplayName":"Text Embedding","taskType":"embedding","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":512,"modelDisplayName":"all-minilm-l6-v2","parameters":"22.7M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md","use_cases":["Semantic Search","Similarity Matching","Document Retrieval","RAG"],"description":"all-MiniLM-L6-v2 is a lightweight sentence embedding model that converts sentences and short paragraphs into 384-dimensional dense vectors. These embeddings capture semantic meaning, making the model well suited for semantic search, retrieval, similarity scoring, clustering, deduplication, recommendation, and text ranking. Its compact architecture is optimized for fast, low-cost inference, making it a strong choice for high-volume embedding workloads and retrieval pipelines.","favicon_url":"https://models-favicon.zerogpu.ai/all-minilm-l6-v2/all-minilm-l6-v2.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/models/all-minilm-l6-v2","privacy_service":"https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md","audio_per_minute":0,"input_per_1m_tokens":0.004,"output_per_1m_tokens":0,"sample_responses_body":{"input":"Apple announced the iPhone 18 Pro and iPhone 18 Pro Max in September 2026, introducing its latest generation of premium iPhones alongside new Apple Intelligence features and updated hardware designed for AI-powered experiences.","model":"all-minilm-l6-v2"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":{"input":"Apple is expected to unveil its next-generation M5 chip at WWDC this June, promising a 40% boost in GPU performance and a new dedicated AI core for on-device machine learning tasks.","model":"all-minilm-l6-v2"}},"displayPriority":0,"createdAt":"2026-08-11T14:55:41.912882+00:00"},{"modelId":"bge-small-en-v1.5","modelType":"BERT","modelVersion":"v1.5","taskDisplayName":"Text Embedding","taskType":"embedding","onDeviceInference":false,"quantized":false,"quantization":null,"maxTokens":512,"modelDisplayName":"bge-small-en-v1.5","parameters":"33M","modelUsecases":null,"pricing":{"currency":"USD","terms_url":"https://huggingface.co/BAAI/bge-small-en-v1.5","use_cases":["Semantic Search","RAG","Similarity Matching","Document Retrieval"],"description":"bge-small-en-v1.5 is a lightweight English embedding model from BAAI designed for semantic search, retrieval, similarity scoring, ranking, and RAG workflows. It converts text into 384-dimensional dense embeddings that capture semantic meaning, making it well suited for finding related documents, matching queries to content, clustering text, and powering vector search. Its compact size makes it a strong choice for high-volume embedding workloads where latency and cost matter.","favicon_url":"https://cdn-avatars.huggingface.co/v1/production/uploads/1664511063789-632c234f42c386ebd2710434.png","model_doc_url":"https://docs.zerogpu.ai/api-reference/models/bge-small-en-v1-5","privacy_service":"https://huggingface.co/BAAI/bge-small-en-v1.5","audio_per_minute":0,"input_per_1m_tokens":0.004,"output_per_1m_tokens":0,"sample_responses_body":{"input":"Apple announced the iPhone 18 Pro and iPhone 18 Pro Max in September 2026, introducing its latest generation of premium iPhones alongside new Apple Intelligence features and updated hardware designed for AI-powered experiences.","model":"bge-small-en-v1.5"},"cached_input_per_1m_tokens":0,"sample_chat_completions_body":null},"displayPriority":0,"createdAt":"2026-08-20T16:18:32.672409+00:00"}]}