LoginSign Up
← Back to all sparks
T

Together AI

AI-ASSISTANTS
Velocity5.5

Open-source AI cloud platform for training and inference of generative models

www.together.ai ↗

Together AI is pricing itself as the open-stack alternative to frontier coding-agent APIs.

inference-economicscoding-agentsopen-modelsdeepseekhuggingfaceday-zero-launches
◆Current state
Together is hammering on two things: (a) inference economics, with a benchmark claiming 76% lower cost than Claude Opus 4.6 on coding-agent workloads, and (b) breadth of model surface, evidenced by day-0 Nemotron 3 Nano Omni, DeepSeek-V4 Pro at 512K context, and Goose-driven 'deploy any HuggingFace model' tooling. Side outputs — a voice finder, the Violin video-translation tool, and a Pearl Research Labs crypto-inference partnership — broaden the developer surface without changing the core narrative.
◆Where it's heading
Together is positioning to be the default API for teams running coding agents on open models, with explicit price/perf comparisons against closed labs. The pattern of day-0 launches plus dedicated container offerings makes the strategy clear: any open frontier model should be one click away on Together. Crypto-adjacent and partnership work (Pearl, Adaption) reads as experimentation rather than core roadmap.
◆Prediction
Expect more cost-comparison content against named frontier APIs and a tighter coding-agent SKU (likely a benchmark-grounded preset for Cursor/Aider-style workloads). Day-0 launch cadence will continue as the differentiator versus AWS Bedrock and other neoclouds.

◆Recent moves

  1. 2d ago

    Benchmarking inference at scale: coding agents

    ⚡ SPARK

    Together publishes head-to-head inference benchmarks framing itself as 76% cheaper than Claude Opus 4.6 on coding agents, with TPS and TTFT advantages against TensorRT-LLM. The most explicit competitive positioning move in the window.

    View source ↗
  2. 6d ago

    Together AI and Pearl Research Labs Team Up to Reduce the Cost of AI Inference

    Partnership with Pearl Research Labs to ship a discounted endpoint for Gemma-4-31B-it-pearl using Proof of Useful Work crypto emissions. Niche and experimental, but a notable cross of crypto and inference economics.

    View source ↗
  3. 7d ago

    Violin: An open-source video translation skill that breaks language barriers

    Violin, an open-source video translation pipeline combining ASR, LLM translation, and TTS. Broadens Together's reference-implementation library and showcases its TTS stack.

    View source ↗
  4. 9d ago

    Introducing voice finder — a new tool to quickly find the right voice for your app from over 600+ voices

    Voice Finder lets developers search 600+ TTS voices using natural-language prompts or audio samples. A UX-layer addition that makes the voice catalog actually navigable.

    View source ↗
  5. 10d ago

    Serving DeepSeek-V4: why million-token context is an inference systems problem

    Deep engineering blog on serving DeepSeek-V4 at million-token context on HGX B200. Useful proof of capability but informational rather than a release.

    View source ↗
  6. 13d ago

    Deploy and inference any model from HuggingFace

    Tutorial-as-product: deploy any Hugging Face model via Goose into Together's Dedicated Container Inference in one prompt. Lowers the barrier between an open model release and a production endpoint on Together.

    View source ↗