@togethercompute profile picture
Together AI @togethercompute

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

Joined November 2022
  • Posts 2,891
  • Following 375
  • Followers 58,021
852 Photos and videos
Pinned Post
Together AI
@togethercompute
Jul 1
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
Jul 1
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! shorturl.at/SooOP
8
20
96
14,437
"Alex really hit the nail on the head, by sending data to a company that has extremely smart models, you are really giving up your business's recipe for them to copy." Our CEO @vipulved when asked about Alex Karp's comments on closed model providers training on YOUR data.
0:35
Palantir
@PalantirTech
Jul 1
Palantir CEO Alex Karp on what customers actually want, the real business of frontier labs, and the importance of open source models: “What the technical customers want is control over their compute, their models, their data stack, and their alpha. They want to know they own the means of production, and it's not being transferred to someone else.” "Who owns the data? Are the prompts secure? Is this being transferred to you?" "If it was so valuable, and I can make you a billion dollars, wouldn't I say I'll make you a billion dollars and I want 30%? Why are they charging for tokens if it's so valuable?"
2:49
1
4
20
6,079
Open-source models give you complete control, customization, and ownership over your data. Companies are moving fast on this. @vipulved on @CNBC with @dee_bosa
0:37
3
4
39
16,143
We're releasing the full slides for our 2 hr deepdive session from the AI Engineer World's Fair. We covered how we build inference engines to serve agentic workloads at trillion token production scale. Slides ⬇️
Zain
@ZainHasan6
Jul 3
Full slides for our 2 hr inference engines deepdive from AI Engineer World's Fair 2026 👇 We cover: > the lifetime of a request > how the enginecore works > how GPU workers function > parallelism configs > spec decoding
5
6
77
14,245
Open model usage has gone from 10% of AI tokens to 30% in a year. The shift to open, modular AI is here to stay. Our founders on what's driving it, and where it goes from here.
0:35
3
8
82
7,866
Read more on the economy of tokens by @vipulved x.com/vipulved/status/207140…
Vipul Ved Prakash
@vipulved
Jun 29
x.com/i/article/207135784544…
3
1,388
We analyzed GLM 5.2 vs Sonnet 5 for software engineering tasks using DeepSWE. GLM 5.2 gets you ~80% of Sonnet 5's capability at ~20% of the price. More insights in the thread!
Zain
@ZainHasan6
Jul 2
Deepdive: Sonnet 5 and GLM-5.2 on DeepSWE > both max reasoning effort > on full DeepSWE benchmark 113 original long-horizon SWE tasks > 4 trials each This is the what the trajectories show 👇(1/n)🧵
14
19
375
51,843
Our CEO @vipulved on @CNBC with @dee_bosa: your data is your recipe. As models get smarter, sending proprietary workflows, customer context, and business logic into closed systems becomes a strategic decision. Open models help companies build AI while keeping more of their intelligence layer under their own control.
0:35
4
9
42
4,798
Together AI reposted
Deirdre Bosa
@dee_bosa
Jul 2
LIVE at 12p PT/3p ET: AI’s cleanest stories are getting messy. Meta has excess compute. AI-heavy companies are hiring, not shrinking. And open models keep getting better. @DanielTNiles on what this means for the AI trade. @arakharazian and @garrettlord on the jobs data. @vipulved and @ClementDelangue on open models, China, and where the money goes if AI keeps getting cheaper.
5
9
33
4,769
30 billion tokens a month to 400 trillion in a year. That's @cursor_ai, @DecagonAI, @cartesia and hundreds of other teams choosing open infrastructure. Our Series C is fuel to take it further.
2
52
4,426
See how our latest funding round accelerates the shift to open-source AI: together.ai/blog/announcing-…

Announcing our $800M Series C to accelerate the shift to open-source AI

We raised $800M to accelerate the shift to open-source AI. Here's why the economics of closed models don't scale, and what we're building next.

together.ai
1
1
8
1,004
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.
2
30
4,034
Under the hood: 6x cost reduction per turn, p95 latency under 400ms, and models shipping weekly. together.ai/customers/decago…

How Decagon Engineered Sub-Second Voice AI with Together AI

Decagon partnered with Together AI to deliver sub-second voice AI with NVIDIA Blackwell GPUs, achieving <400ms latency, 6Ă— cost reduction, and a conversational experience customers thank the agent...

together.ai
4
1,246
Read more about why we built this and where we're going. together.ai/blog/announcing-…

Announcing our $800M Series C to accelerate the shift to open-source AI

We raised $800M to accelerate the shift to open-source AI. Here's why the economics of closed models don't scale, and what we're building next.

together.ai
3
816
Intelligence should be abundant, not expensive.
2:35
5
10
68
5,893
Read about our Series C in the New York Times: shorturl.at/SooOP

As Companies Race for Cheaper A.I. Options, This Start-Up Pitches a Solution (Gift Article)

Together AI, which specializes in open-source artificial intelligence models, is now worth more than $8 billion.

nytimes.com
1
8
1,369
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 available on Together AI.
5
16
97
13,045
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 the Together AI Frontier Performance team.
8
30
250
16,513
Read more on ParallelKernelBench: together.ai/blog/parallelker…

ParallelKernelBench: Frontier LLMs can't write fast multi-GPU kernels (yet)

ParallelKernelBench tests whether LLMs can write fast multi-GPU CUDA kernels across 87 real workloads. The best model solves under a third, but a few generated kernels beat any public implementation.

together.ai
1
1
15
1,575
7/ Untied Ulysses: Memory-Efficient Context Parallelism via Headwise Chunking Paper: arxiv.org/abs/2602.21196
1
1
949
8/ Opportunistic Expert Activation: Batch-Aware Expert Routing for Faster Decode Without Retraining (OEA) Paper: arxiv.org/abs/2511.02237
1
1
949
9/ ParallelKernelBench: Benchmarking LLMs on Multi-GPU Kernel Generation Paper: alphaxiv.org/abs/2606.parall…

ParallelKernelBench: Can LLMs Write Fast Multi-GPU Kernels?

ParallelKernelBench (PKB), a new benchmark and evaluation framework, assesses the ability of large language models (LLMs) to generate high-performance multi-GPU kernels. The evaluation shows...

alphaxiv.org
1
1
8
1,297
Load more