Together AI (@togethercompute) / X

Together AI

2,880 posts

Together AI

@togethercompute

Accelerate inference, model shaping, and pre-training on a research-optimized platform.

San Francisco, CA

Joined November 2022

Pinned
Together AI
@togethercompute
12h
Four years ago we made a bet on open-source AI. Today the numbers speak for themselves. Proud to announce our next milestone🚀
Vipul Ved Prakash
@vipulved
12h
We @togethercompute believe intelligence should be abundant, not expensive. Today we announced our Series C funding of $800m @ $8.3B valuation, to continue to build the world's most efficient platform for generative AI. Thanks @nikogallogly for telling our story in @nytimes!
9K
Together AI
@togethercompute
2h
Tasks moved from closed models to open models on Together AI cost one-fifth to one-seventh as much. That's from @DecagonAI co-founder @AshwinSreenivas in the NYT today. This is what abundant intelligence looks like in practice.
1.5K
Together AI
@togethercompute
2h
Under the hood: 6x cost reduction per turn, p95 latency under 400ms, and models shipping weekly.
How Decagon Engineered Sub-Second Voice AI with Together AI
From together.ai
495
Together AI
@togethercompute
2h
Replying to @togethercompute
Read more about why we built this and where we're going.
Announcing our $800M Series C to accelerate the shift to open-source AI
From together.ai
346
Together AI
@togethercompute
10h
Replying to @togethercompute
Read about our Series C in the New York Times:
As Companies Race for Cheaper A.I. Options, This Start-Up Pitches a Solution (Gift Article)
From nytimes.com
958
Together AI
@togethercompute
10h
Intelligence should be abundant, not expensive.
00:00
4.4K
Together AI
@togethercompute
20h
Nemotron 3 Ultra is taking off on Together AI 🚀 In just days, it's climbed to 35B tokens/day on @OpenRouter. That's the community voting with their tokens for open models that are fast, efficient, and fully customizable. A top open model from @nvidia, fully customizable and
8K
Together AI
@togethercompute
21h
Replying to @togethercompute
Read more on ParallelKernelBench:
ParallelKernelBench: Frontier LLMs can't write fast multi-GPU kernels (yet)
From together.ai
1.2K
Together AI
@togethercompute
21h
Multi-GPU kernels are the real test for coding models. Today at @aiDotEngineer, @simran_s_arora shared ParallelKernelBench, an open-source benchmark for evaluating whether LLMs can write fast CUDA kernels for real communication-heavy workloads. Proud to see this work from
15K
Together AI
@togethercompute
22h
Replying to @togethercompute
9/ ParallelKernelBench: Benchmarking LLMs on Multi-GPU Kernel Generation Paper:
ParallelKernelBench: Can LLMs Write Fast Multi-GPU Kernels?
From alphaxiv.org
1.1K
Together AI
@togethercompute
22h
Replying to @togethercompute
8/ Opportunistic Expert Activation: Batch-Aware Expert Routing for Faster Decode Without Retraining (OEA) Paper: arxiv.org/abs/2511.02237
786
Together AI
@togethercompute
22h
Replying to @togethercompute
7/ Untied Ulysses: Memory-Efficient Context Parallelism via Headwise Chunking Paper: arxiv.org/abs/2602.21196
784
Together AI
@togethercompute
22h
Replying to @togethercompute
6/ When RL Meets Adaptive Speculative Training: A Unified Training-Serving System (Aurora) Paper: arxiv.org/abs/2602.06932
511
Together AI
@togethercompute
22h
Replying to @togethercompute
5/ V1: Unifying Generation and Self-Verification for Parallel Reasoners Paper: arxiv.org/abs/2603.04304
552