Qwen4 Max is live on Together AI
Together AI made Alibaba's flagship Qwen4 Max available on Together Serverless Inference with a 1M-token context window, and in a test the model took a bug report plus all 344 files of the Tornado Pyt…
Together AI made Alibaba's flagship Qwen4 Max available on Together Serverless Inference with a 1M-token context window, and in a test the model took a bug report plus all 344 files of the Tornado Pyt…
Together AI disclosed that an unauthorized party downloaded stored prompt samples from its serverless inference service submitted in June and July 2026, after an AI coding agent committed a credential…
Together AI began serving Qwen's open text-to-speech model Qwen3 TTS on its Serverless Inference and dedicated endpoints at $20 per 1M characters, with first audio in as little as 97ms. The model offe…
Together AI launched Together Storage Accelerator (TSA), a distributed caching layer included free with every Together GPU Cluster that caches object-storage and Hugging Face Hub reads on cluster NVMe…
Together AI began serving Tencent's flagship Hy4 model through serverless and dedicated endpoints under the model string tencent/Hy4, priced at $0.83 per 1M input tokens, $2.50 per 1M output tokens an…
Together AI published an account of how it built agent evaluations to measure whether coding agents can use its platform, running each task in three configurations — a baseline with no special setup, …
Together AI recovered 98% of megakernel inference performance using only CUDA Graphs and Programmatic Dependent Launch (PDL) on top of default Torch Inductor kernels, without writing a single custom k…
Together AI and Ollama published a joint write-up on the client-side and inference-engine-side requirements for keeping prefix caching intact in long-horizon agentic workloads, where coding agents res…
Together AI introduced Together Compiler (TKCC), a continual optimization loop that uses high-throughput agents to tune inference workloads, achieving up to 2.6x higher TPS across frontier open-source…
Together AI introduced Together Compiler (TKCC), a continual agentic system that optimizes AI inference workloads, reaching 1.3-3x higher TPS across frontier open-source models including Kimi K3, GLM …
Together AI has made its Automatic node remediation system generally available for its GPU clusters, with the company's GPU cluster customers and internal inference platform already using it to reduce…
Together AI launched Together Post-Training, a service that lets developers build custom models on its platform through reinforcement learning from environment rewards and supervised fine-tuning on th…
Meta's Muse Spark Open, described by Together AI as the most capable US open-weight model to date, became available on Together AI on its launch day as a Day 0 partner release. The model offers a 1M-t…
Together AI launched Together Link, a tool that connects existing coding-agent harnesses — Claude Code, Claude Desktop, Codex in the ChatGPT app and CLI, OpenCode and Pi — to open models on Together A…
A developer launched MuseTogether, a free service that lets Meta Muse agents exchange messages with each other so they can coordinate plans such as dinners, family visits and weekends on behalf of the…
Together AI launched together/Tev1-4B-experimental, a Jev-like classification model built on Qwen3.5 4B and served on Together's serverless platform, and published a tutorial showing how to fine-tune …
A global fintech moved its AI coding-assistant workload to GLM-5.2 served through Together's Dedicated Model Inference, giving its engineers self-service endpoint scaling instead of capacity requests …
Together published a migration guide for companies moving from closed-source to open-source AI models, arguing the process can take weeks to months rather than months to years when a managed service i…
Francesco launched ProveTogether, a Show HN project that lets users connect AI agents to collaboratively discuss math conjectures and submit proofs in Lean, with verified proof source code made public…
Together AI expanded its Together Fine-Tuning service to support 17 new open-weight models, including GLM 5.3, DeepSeek-V4-Flash, Kimi K2.7-Code, Qwen 3.5 and Gemma 4, alongside live experiment tracki…