Together · Blogs

Blogs listings from Together on TechiSeek. 208 public listings.

• RedPajama 7B now available, instruct model outperforms all open 7B models on HELM benchmarks (Together)
  The Red Pajama project aims to create a set of leading open-source models and to rigorously understand the ingredients that yield good perf…
• FAQ: Building LLMs with RedPajama-v2, a 30 trillion token web dataset (Together)
  Over the past several months, we have been amazed by the community's engagement with the Red Pajama-V2 dataset.
• RedPajama-INCITE-3B, an LLM for everyone (Together)
  We are excited to share a set of updates that make it even easier to use and fine-tune Red Pajama-INCITE-3 B, including Red Pajama support…
• ThunderAgent: 2x Faster Agentic Inference for Synthetic Data Generation at Scale (Together)
  We introduce Thunder Agent, a system for high throughput agentic inference. By introducing a novel program abstraction for agentic LLM requ…
• TEAL: Training-Free Activation Sparsity in Large Language Models (Together)
  We present TEAL (Training-Free Activation Sparsity in LLMs), a simple training-free approach to activation sparsity that applies magnitude…
• Releasing 3B and 7B RedPajama-INCITE family of models including base, instruction-tuned & chat (Together)
  The Red Pajama project aims to create a set of leading open-source models and to rigorously understand the ingredients that yield good perf…
• The production platform for open-weight AI inference (Together)
  We're releasing a significant update to our inference platform, giving you complete control over performance, cost, and quality without bui…
• Generate images with specific styles using Flux LoRAs on Together AI (Together)
  Flux LoRAs are a really powerful way to customize AI image generation for specific styles.
• Flash-Decoding for long-context inference (Together)
  Large language models (LLM) such as ChatGPT or Llama have received unprecedented attention lately.
• The Frontier is Open (Together)
  Peter Thiel once said that good startups are built around secrets: ideas that only a few people see at first but that becomes obvious later.
• Expanding our enterprise inference capacity with IBM Cloud and NVIDIA (Together)
  We’re excited to share that Together AI is working with IBM and NVIDIA to scale enterprise-grade AI inference, starting with a large cluste…
• Paving the way to efficient architectures: StripedHyena-7B, open source models offering a glimpse (Together)
  One of the focus areas at Together Research is new architectures for long context, improved training, and inference performance over the Tr…
• Building your own RAG application using Together AI and MongoDB Atlas (Together)
  Together AI provides the fastest cloud platform for building and running generative AI. As part of a series of blog posts about the Togethe…
• RedPajama, a project to create leading open-source models, starts by reproducing LLaMA training (Together)
  Foundation models such as GPT-4 have driven rapid improvement in AI. However, the most powerful models are closed commercial models or only…
• Together AI delivers fastest inference for the top open-source models (Together)
  Together AI now delivers up to 2 x faster serverless inference for demanding open-source LLMs, ranking #1 in output speed benchmarks.
• Sequoia: Scalable, Robust, and Hardware-aware Speculative Decoding (Together)
  We introduce Sequoia, a scalable, robust, and hardware-aware speculative decoding framework that improves LLM inference speed on consumer G…
• How to deploy DeepSeek-R1 and distilled models securely on Together AI (Together)
  Deep Seek-R1 has taken the world by storm by establishing itself as a formidable open-weight competitor to proprietary reasoning models lik…
• How to Build a State-of-the-Art Search Stack for LLMs: RAG, Reranking, and Reinforcement Learning (Together)
  As AI systems increasingly rely on external context to generate accurate, helpful responses, high-performance search infrastructure is no l…
• SpecExec: Massively Parallel Speculative Decoding for Interactive LLM Inference on Consumer Devices (Together)
  We introduce Spec Exec, a new speculative decoding method that applies the classical approach of “speculative execution” to LLM inference.
• RedPajama training progress at 440 billion tokens (Together)
  A week ago we announced Red Pajama, a project to create leading open-source models. We released the first step in the project a training da…
• Together AI and Snorkel AI empower enterprises to build proprietary LLMs (Together)
  We are proud to announce a new strategic partnership with Snorkel AI to enable organizations to build custom LLMs on their data in their se…
• Transform OpenAI gpt-oss Models into Domain Experts with Together AI Fine-Tuning (Together)
  Transform OpenAI gpt-oss Models into Domain Experts with Together AI Fine-Tuning The release of OpenAI's gpt-oss-120 B and gpt-oss-20 B mod…
• Even Better, Even Faster Quantized LLMs with QTIP (Together)
  We’re excited to announce QTIP (Quantization with Trellises and Incoherence Processing), a new weight-only LLM post-training quantization m…
• Fine-tuning language models over slow networks using activation compression with guarantees (Together)
  Communication compression is a crucial technique for modern distributed learning systems to alleviate their communication bottlenecks over…
• EinsteinArena: Harnessing the collective intelligence of agents in the wild to advance science (Together)
  Scientific discovery has driven human progress, but tackling today’s hardest problems requires collective intelligence beyond any single re…
• FLUX API is now available on Together AI: New FLUX1.1 [pro] and free access to FLUX.1 [schnell] (Together)
  We’re excited to announce support for the FLUX model family, including our fast Together Turbo endpoint for the SOTA open-source FLUX.1 [sc…
• Introducing the Together Embeddings endpoint (Together)
  Today, we are excited to release the Together Embeddings endpoint!
• FLUX.1 Kontext models: Character consistency and precise image editing without fine-tuning (Together)
  Starting today on Together AI, you can access FLUX.1 Kontext [max] and FLUX.1 Kontext [pro] from Black Forest Labs (BFL), a suite of genera…
• Flash Attention received the inaugural Stanford open source software award (Together)
  Flash Attention, invented by Tri Dao (Chief Scientist of Together AI), Dan Fu (Academic partner at Together AI), Stefano Ermon (Advisor at…
• RedPajama-Data-v2: An open dataset with 30 trillion tokens for training large language models (Together)
  Today, we're releasing a new version of the Red Pajama dataset, with 30 trillion filtered and deduplicated tokens (100+ trillions raw) from…
• Releasing GPT-JT powered by open-source AI (Together)
  A fork of GPT-J-6 B, fine-tuned on 3.53 billion tokens, that outperforms most 100 B+ parameter models at classification.
• Evo: Long-context modeling from molecular to genome scale (Together)
  Introducing Evo, a long-context biological foundation model based on the Striped Hyena architecture that generalizes across the fundamental…
• How to evaluate and benchmark Large Language Models (LLMs) (Together)
  Learn how to evaluate and benchmark large language models using datasets like MMLU, GSM8 K, and Human Eval.
• GLM-5.3 vs. Claude Fable 5 on DeepSWE: Cost, Coding, and Routing (Together)
  GLM-5.3 and Claude Fable 5 finish within noise of each other on DeepSWE accuracy, but GLM-5.3 costs a fifth as much per task and wins every…
• Serving DeepSeek-V4: why million-token context is an inference systems problem (Together)
  Benchmark tables miss the main point of Deep Seek-V4: the important change is architectural.
• Accelerate RL rollouts by up to 50% with distribution-aware speculative decoding (Together)
  Distribution-aware speculative decoding (DAS) is a novel framework that significantly alleviates the rollout bottleneck in RL post-training…
• Together AI Delivers Top Speeds for DeepSeek-R1-0528 Inference on NVIDIA Blackwell (Together)
  Earlier this year, we invited select customers—including Zoom, Salesforce, and In Video—to test-drive NVIDIA Blackwell GPUs on Together GPU…
• Direct Preference Optimization: A Technical Deep Dive (Together)
  We're excited to announce that the Together Fine-Tuning Platform now supports Direct Preference Optimization (DPO)!
• Dragonfly: A large vision-language model with multi-resolution zoom (Together)
  We are excited to announce the launch of Dragonfly, an instruction-tuning Vision-language architecture, that enhances fine-grained visual u…
• Fine-Tuning LLMs for Multi-Turn Conversations: A Technical Deep Dive (Together)
  Large Language Models (LLMs) have revolutionized how we interact with and build conversational AI systems.
• Powering Secure AI: Together AI Achieves SOC 2 Type 2 Compliance (Together)
  We are excited to announce that Together AI has successfully completed a SOC 2 Type 2 examination.
• The Mamba in the Llama: Distilling and Accelerating Hybrid Models (Together)
  The evolution of large language models (LLMs) has been largely driven by the success of Transformer architectures.
• Together AI expands fine-tuning service with tool calling, reasoning, and vision support (Together)
  - Tool call fine-tuning: Ensure agents execute structured actions reliably with end-to-end fine-tuning and inference on OpenAI-compatible s…
• Announcing Together Custom Models. Build a state-of-the-art LLM with Together AI (Together)
  Training your own state-of-the-art LLM enables you to achieve the highest accuracy and adaptability to your tasks, with the best price-perf…
• What does 99.9% uptime mean for inference? (Together)
  - The short version: Each reliability tier maps to a specific failure domain, and each one requires its own architecture to survive it.
• Together AI launches full stack for developers to build with open-source AI (Together)
  Introducing Together API and Together Compute — simple, powerful, cost effective cloud services to train, fine-tune, and run the world’s le…
• How Together AI Uses AI Agents to Automate Complex Engineering Tasks: Lessons from Developing (Together)
  TLDR: Building AI agents to handle complex and long-running engineering tasks requires a different approach than typical AI agent applicati…
• Announcing Together AI and Adaption Partnership (Together)
  We’re excited to partner with Adaption to make Together Fine-Tuning available in Adaptive Data.
• Growing to 20 exaflops, Together GPU Clusters help startups and enterprises accelerate generative AI (Together)
  Generative AI has taken the world by storm, and its capabilities are only increasing. Venture capital funding for AI startups has more than…
• Together AI welcomes Alon Gavrielov as VP of Infrastructure Strategy (Together)
  The rapid acceleration in AI capabilities is fueling a wave of AI-native applications that are transforming experiences across domains—from…