Kimi K3 matches Claude Fable 5 on DeepSWE benchmark quality with a 68.5% pass@1 versus Fable's 69.9%, but costs $4.65 per rollout compared to Fable's $13.41, delivering 2.8x more solved tasks per doll…
Together AI released a significant update to its inference platform, giving users complete control over performance, cost, and quality without building their own stack. The platform supports open-weig…
Together AI and Y Combinator have partnered to launch the first dedicated GPU cluster for Y Combinator's portfolio of AI-native startups, providing flexible, short-term access to compute at long-term …
Together, which runs inference for Cursor, Decagon, Cartesia, and Yutori, explains that 99.9% inference uptime requires surviving a full data center failure through multi-DC deployment with live traff…
Together AI has shipped a set of changes to Together GPU Clusters focused on platform health and operational control, including passive health checks, auto node repair, and Together Slurm-on-K8s 2.0. …
Thinking Machines Lab released Inkling, a new multimodal mixture-of-experts model for token-efficient reasoning and native multimodal understanding, and Together AI made it available on its inference …
Together AI launched Provisioned Throughput, a new inference service offering reserved capacity for open-weight models with token-based pricing and a 99% uptime SLA, available for MiniMax M3 and GLM-5…
Together AI announced an $800 million Series C funding round led by Aramco Ventures, NVIDIA, and others to accelerate open-source AI development. The company plans to use the capital to expand compute…
Together AI presented eight papers at ICML 2026 spanning the full AI stack, from frontier agents to GPU kernels. Their research includes DSGym for standardized data-science agent evaluation, ThunderAg…
Frontier LLMs including GPT-5.5, Gemini 3 Pro, and Opus 4.7 failed to correctly solve over two-thirds of problems in ParallelKernelBench, a new benchmark for multi-GPU kernel generation. The benchmark…
An experiment comparing Kimi K2.7 Code and Claude Fable 5 across 12 landing pages found that Kimi cost 94% less while achieving similar quality scores. The open-source model was roughly 16 times cheap…
Together AI has earned ISO 27001:2022 certification from A-LIGN Compliance and Security, Inc., confirming its Information Security Management System meets the latest international standard for informa…
Together AI partnered with MiniMax to serve the new M3 model in production, achieving 81–125% throughput improvements through engineering optimizations including a KV-Block-Major sparse attention kern…
Together AI built the world’s fastest speech-to-text stack, enabling NVIDIA’s Parakeet-TDT 0.6B v3 model to transcribe roughly 20 hours of speech in under 10 seconds. The company achieved this by opti…
Together Inference Engine delivered 31% more tokens per second than the next fastest open-source inference engine on the same hardware during a production coding agent workload, while maintaining twic…
Together AI and Pearl Research Labs have partnered to launch a new AI inference endpoint for the Gemma-4-31B-it-pearl model, offering over 25% discounts by offsetting costs through cryptocurrency mini…
Together AI has released Violin, an open-source video translation tool that uses speech recognition, large language models, and speech synthesis to convert video content across languages. The tool, po…
Together AI launched Voice Finder, a search tool that indexes over 600 voices across multiple text-to-speech models to help developers quickly select the right voice for their applications. The tool a…
DeepSeek-V4's million-token context capability stems from a hybrid attention architecture that compresses context before KV storage, reducing cache pressure. Together's early bring-up on NVIDIA HGX B2…
Netflix released void-model on Hugging Face, and a developer used the Goose CLI agent with Together's dedicated containers skill to deploy the model for inference on release day with a single prompt. …