Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for local autonomous agentic workflows on consumer hardware. It combines multi-step reasoning, reliable schema-based tool calling, and failure recovery with multimodal understanding via a built-in perception encoder. It supports interleaved text and image input, controllable reasoning strength, and is highly compatible with agentic orchestration frameworks.

NVIDIA Nemotron-3 Ultra 550B A55B is a frontier-scale large language model from NVIDIA, designed for strong agentic, reasoning, and conversational capabilities. It employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2, MoE, and Attention layers, alongside Multi-Token Prediction (MTP) for faster inference. With a 1-million-token context window and a configurable reasoning mode, it excels at complex autonomous agents, long-context analysis, and deep research workflows.

Qwen3.8-2.4T-A95B is Qwen's largest open-weight Qwen3.8 reasoning model, built for advanced coding, professional work, research, and long-horizon agentic tasks. It uses a Mixture-of-Experts architecture with 2.4 trillion total parameters and 95 billion activated parameters, combining Gated DeltaNet linear attention, gated attention, MoE layers, and multi-token prediction. The model provides a native 262K-token context window that can be extended to approximately 1M tokens, supports configurable reasoning effort, and is optimized for complex multi-step task execution, autonomous planning, coding agents, and long-context workloads.

Tev1-4B-experimental is Together AI's experimental decision model built by supervised fine-tuning Qwen3.5-4B for structured classification and routing tasks. Given an application state, a question, and between 2 and 24 labeled options, the model selects exactly one option rather than producing a conventional free-form response. It is designed for fast decision-making workloads such as intent classification, request routing, policy checks, moderation, workflow automation, and other closed-set decisions. Tev1 is inspired by Jev-style decision models but retains Qwen's standard autoregressive next-token language-model head, allowing it to run through conventional chat-completions infrastructure. Together AI publishes the training recipe and evaluation tooling, making the model useful both as a production decision model and as a reference for building custom classifiers.

GLM-5.3-Flash is Z.AI's natively multimodal Mixture-of-Experts model in the GLM-5 family, designed for high-performance coding, agentic engineering, multimodal reasoning, long-context processing, and efficient inference. It has 320B total parameters with 18B active parameters and introduces a hybrid architecture combining sparse and linear attention together with Manifold-Constrained Hyper-Connections for improved scaling and long-context efficiency. The model supports controllable reasoning effort and is optimized to deliver frontier-level capability at Flash-tier serving cost.

Model	Category	Input	Output	Context	Pricing
Meta: Muse Glimmer 30B e059ca29986448038c397070d82af892	text	text, image	text	131K	$0.350 in / $1.500 out
NVIDIA: Nemotron-3 Ultra 550B A55B cc71cb88b4d54a07a1efc693d77ec72c	text	text	text	1M	$0.600 in / $3.600 out
Qwen: Qwen3.8 2.4T A95B 58b3bc75217343f292950f587092d62f	text	text	text	1M	$2.000 in / $6.000 out
Together AI: Tev1 4B Experimental 6f745bb351e24104a9b03b9403e1868d	text	text	decisions	32K	$0.042 in / $0.000 out
ZAI: GLM 5.3 Flash 17530b712f79494f827aa57b770d65fa	text	text, image	text	1M	$0.150 in / $0.500 out

together_ai models on OneInfer

Categories

Modalities

API access

together_ai model catalog