90
As GPU throughput outpaces memory bandwidth, kernels must evolve. We introduce FlashAttention-4, featuring new pipelining for maximum overlap, 2-CTA MMA modes to reduce shared memory traffic, and a hardware-software hybrid approach to softmax exponentials.
At AI Native Conf, Together AI announced breakthroughs across kernels, RL, and inference optimization — including FlashAttention-4, ThunderAgent, and together.compile. Research that ships to production. That's the AI Native Cloud.
Serving long prompts doesn't have to mean slow responses. Learn how Together AI's CPD architecture separates warm and cold inference workloads to deliver 40% higher throughput and dramatically lower time-to-first-token for long-context LLM serving.
We've refreshed our visual identity — designed with Pentagram to express how Together AI connects open-source innovation, systems research, and builders to unlock new possibilities.
State-of-the-art speech models like Whisper and Deepgram score near-human on benchmarks — then fail 39% of the time on street names. New research from Together AI exposes the gap and a fix.
Standard diffusion language models can't use KV caching and need too many refinement steps to be practical. CDLM fixes both with a post-training recipe that enables exact block-wise KV caching and trajectory-consistent step reduction — delivering up to 14.5x latency improvements
Together AI launches production-grade orchestration for custom AI models with 1.4x–2.6x faster inference.
What do language models generate when you don't tell them what to generate? New research reveals that LLM families have distinct 'knowledge priors'—GPT models default to code and math, Llama favors narratives, DeepSeek generates religious content, and Qwen outputs exam questions.
Hiring Alon Gavrielov further deepens Together AI’s commitment to building AI factories that deliver the most reliable, efficient, and scalable infrastructure for AI-native teams.
Together Evaluations now supports OpenAI, Anthropic, and Google models for cross-provider benchmarking. Compare open-source, fine-tuned, and proprietary models side-by-side to make data-driven decisions on quality, cost, and performance—all in one platform.
Fine-tuned open-source LLM judges can outperform GPT-5.2 at evaluating model outputs. Using Direct Preference Optimization on just 5,400 preference pairs, we trained GPT-OSS 120B to beat GPT-5.2 on human preference alignment—at 15x lower cost and 14x faster inference speeds.
Introducing DSGym—a holisti evaluation and training framework for LLM-based data science agents. Features 90+ bioinformatics tasks, 92 Kaggle competitions, and synthetic trajectory generation. Our 4B model achieves state-of-the-art performance among open-source models through exe
Learn how to reduce inference latency without massive cost using proven inference optimization tactics — improving throughput, GPU utilization, and cost efficiency while balancing throughput vs. latency tradeoffs.
Together AI teamed with Cursor to build the real-time inference stack that keeps in-editor agents fast and reliable. They productionized NVIDIA Blackwell (B200/GB200), tuning ARM hosts, kernels, and FP4/TensorRT quantization for low latency and rapid model rollouts.
Learn how foundation models are trained at scale using multi-node GPU clusters, including distributed training techniques, infrastructure requirements, and practical steps to scale training efficiently.
Learn how to choose the right open-source model for production by evaluating model quality, benchmarking performance, and deploying open models that balance cost, speed, and accuracy.
MiniMax Speech 2.6 Turbo: State-of-the-art multilingual TTS with human-level emotional awareness, sub-250ms latency, and 40+ languages—now on Together AI.
Two enterprise-grade Rime TTS models now available on Together AI. Co-locate with LLM and STT on dedicated infrastructure. Proven at billions of calls.
Dan Fu, our VP of Kernels, has published a new post challenging the idea that AI is hitting a hardware wall. He argues that we are vastly underutilizing current chips and that better software-hardware co-design will unlock the next order of magnitude in performance.
Nemotron 3 Nano, NVIDIA’s newest reasoning model, is now available on Together AI, the AI Native Cloud
