{"object":"list","data":[{"modelId":"zai-org/GLM-5.3-Flash","name":"GLM 5.3-Flash","organization":"ZAI","description":"GLM-5.3-Flash is the first natively multimodal model in Z.ai's GLM-5 series, a 320B-parameter Mixture-of-Experts design that activates 18B per token. A hybrid attention architecture pairs linear attention for local context with sparse attention and a lightweight indexer for global retrieval, cutting long-context serving cost across its 1M-token window. Trained on a 30T-token multimodal corpus, it reads images, video, and documents alongside code and lands close to flagship-class results on agentic coding benchmarks at a fraction of the price.","tagline":"Natively multimodal coding model with 1M context at flash-tier pricing","license":"MIT","linkToHuggingFace":"https://huggingface.co/zai-org/GLM-5.3-Flash","releaseDate":"2026-08-26","architecture":"Transformer","isMixtureOfExperts":true,"totalParameters":320000000000,"activatedParameters":18000000000,"contextLength":1048576,"maxOutputTokens":131072,"quantizationLevel":null,"type":["Chat","Code","Vision","LLM"],"mainUseCase":["Code","Vision"],"inputModality":["Text","Image"],"outputModality":["Text"],"deployment":["Serverless"],"speed":"High","intelligence":"High","inputPricePerMillionTokens":0.15,"cachedInputPricePerMillionTokens":0.03,"outputPricePerMillionTokens":0.25,"audioPricePerMillionUtf8Bytes":null,"fineTuning":false,"functionCalling":true,"embeddings":false,"rerankers":false,"id":"zai-org/GLM-5.3-Flash","object":"model","created":1787702400,"owned_by":"ZAI"},{"modelId":"zai-org/GLM-5.3-Flash-FP4","name":"GLM-5.3-Flash-FP4","organization":"ZAI","description":"GLM-5.3-Flash-FP4 is GLM-5.3-Flash served from an MXFP4-quantized checkpoint. GLM-5.3-Flash is the first natively multimodal model in Z.ai's GLM-5 series, a 320B-parameter Mixture-of-Experts design that activates 18B per token, pairing linear attention for local context with sparse attention and a lightweight indexer for global retrieval across its 1M-token window. The 4-bit expert weights cut serving cost while staying close to the FP8 model on agentic coding benchmarks.","tagline":"GLM-5.3-Flash in MXFP4: 1M context and image input at a lower price","license":"MIT","linkToHuggingFace":"https://huggingface.co/amd/GLM-5.3-Flash-Quark-MXFP4","releaseDate":"2026-11-12","architecture":"Transformer","isMixtureOfExperts":true,"totalParameters":320000000000,"activatedParameters":18000000000,"contextLength":1048576,"maxOutputTokens":131072,"quantizationLevel":"MXFP4","type":["Chat","Code","Vision","LLM"],"mainUseCase":["Code","Vision"],"inputModality":["Text","Image"],"outputModality":["Text"],"deployment":["Serverless"],"speed":"High","intelligence":"High","inputPricePerMillionTokens":0.1125,"cachedInputPricePerMillionTokens":0.0225,"outputPricePerMillionTokens":0.1875,"audioPricePerMillionUtf8Bytes":null,"fineTuning":false,"functionCalling":true,"embeddings":false,"rerankers":false,"id":"zai-org/GLM-5.3-Flash-FP4","object":"model","created":1794441600,"owned_by":"ZAI"},{"modelId":"zyphra/ZAYA1-8B","name":"ZAYA1-8B","organization":"Zyphra","description":"ZAYA1-8B is an open-source small MoE model by Zyphra that pushes the frontier of intelligence density through its novel architecture and unique test-time compute inference. ZAYA1-8B excels at challenging mathematical and coding tasks that require deep reasoning expertise.","tagline":"Efficient small language model for mathematical and coding problems.","license":"Apache 2.0","linkToHuggingFace":"https://huggingface.co/Zyphra/Zaya1-8B","releaseDate":"2026-05-06","architecture":"Transformer","isMixtureOfExperts":true,"totalParameters":8000000000,"activatedParameters":760000000,"contextLength":131072,"maxOutputTokens":115000,"quantizationLevel":"BF16","type":["Reasoning","LLM"],"mainUseCase":["Reasoning","LLM"],"inputModality":["Text"],"outputModality":["Text"],"deployment":["Serverless"],"speed":"High","intelligence":"Medium","inputPricePerMillionTokens":0.0,"cachedInputPricePerMillionTokens":0.0,"outputPricePerMillionTokens":0.0,"audioPricePerMillionUtf8Bytes":null,"fineTuning":false,"functionCalling":true,"embeddings":false,"rerankers":false,"id":"zyphra/ZAYA1-8B","object":"model","created":1778025600,"owned_by":"Zyphra"},{"modelId":"zyphra/ZONOS2","name":"ZONOS2","organization":"Zyphra","description":"ZONOS2 is an open-source real-time text-to-speech model by Zyphra that delivers high-fidelity voice cloning through its novel MoE architecture, the first in open-source TTS. ZONOS2 leads among open-source TTS models in naturalness and produces life like zero-shot voice clones across a wide range of speakers and languages.","tagline":"Efficient MoE-based TTS with zero-shot voice cloning capabilities.","license":"Apache 2.0","linkToHuggingFace":null,"releaseDate":"2026-05-21","architecture":"Transformer","isMixtureOfExperts":true,"totalParameters":8000000000,"activatedParameters":900000000,"contextLength":6144,"maxOutputTokens":4096,"quantizationLevel":"BF16","type":["Audio"],"mainUseCase":["TTS"],"inputModality":["Text","Audio"],"outputModality":["Audio"],"deployment":["Serverless"],"speed":"High","intelligence":null,"inputPricePerMillionTokens":null,"cachedInputPricePerMillionTokens":null,"outputPricePerMillionTokens":null,"audioPricePerMillionUtf8Bytes":20.0,"fineTuning":false,"functionCalling":false,"embeddings":false,"rerankers":false,"id":"zyphra/ZONOS2","object":"model","created":1779321600,"owned_by":"Zyphra"},{"modelId":"zyphra/ZONOS3-Mini","name":"ZONOS3-Mini","organization":"Zyphra","description":"Open-source text-to-speech with voice cloning, word-level control and real-time performance on-device.","tagline":"High-quality, reliable speech. Just 100M active parameters.","license":"Apache 2.0","linkToHuggingFace":"https://huggingface.co/Zyphra/ZONOS3-Mini","releaseDate":"2026-11-09","architecture":"Transformer","isMixtureOfExperts":true,"totalParameters":886000000,"activatedParameters":100000000,"contextLength":32768,"maxOutputTokens":4096,"quantizationLevel":"BF16","type":["Voice Cloning","Audio"],"mainUseCase":["TTS"],"inputModality":["Text","Audio"],"outputModality":["Audio"],"deployment":["Serverless"],"speed":"Very High","intelligence":null,"inputPricePerMillionTokens":null,"cachedInputPricePerMillionTokens":null,"outputPricePerMillionTokens":null,"audioPricePerMillionUtf8Bytes":10.0,"fineTuning":false,"functionCalling":false,"embeddings":false,"rerankers":false,"id":"zyphra/ZONOS3-Mini","object":"model","created":1794182400,"owned_by":"Zyphra"},{"modelId":"zyphra/ZUNA1.1","name":"ZUNA1.1","organization":"Zyphra","description":"ZUNA1.1 is Zyphra's open, Apache 2.0-licensed EEG foundation model that reconstructs, denoises, and upsamples recordings across arbitrary channel layouts. This 380M-parameter diffusion autoencoder conditions on the true 3D positions of each electrode, working on essentially any montage without retraining and delivering high-fidelity reconstruction across variable-length, real-world recordings.","tagline":"Efficient open foundation model for EEG reconstruction, denoising, and upsampling.","license":"Apache 2.0","linkToHuggingFace":"https://huggingface.co/Zyphra/ZUNA1.1","releaseDate":"2026-07-16","architecture":"Diffusion Autoencoder","isMixtureOfExperts":false,"totalParameters":380000000,"activatedParameters":380000000,"contextLength":7680,"maxOutputTokens":4096,"quantizationLevel":null,"type":["Thought","EEG"],"mainUseCase":["EEG"],"inputModality":["EEG"],"outputModality":["EEG"],"deployment":["Serverless"],"speed":"High","intelligence":null,"inputPricePerMillionTokens":null,"cachedInputPricePerMillionTokens":null,"outputPricePerMillionTokens":null,"audioPricePerMillionUtf8Bytes":null,"fineTuning":false,"functionCalling":false,"embeddings":false,"rerankers":false,"id":"zyphra/ZUNA1.1","object":"model","created":1784160000,"owned_by":"Zyphra"}]}