Together AI (@togethercompute) / X

Together AI

2,791 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
1h
Closed-source models aren't worth the premium. We generated 12 landing pages with Kimi K2.7 Code and Claude Fable 5. Kimi came in 16x cheaper with comparable quality, especially once we gave it visual context through a design MCP server. Open-source models are already a
Hassan
@nutlope
3h
Article
Kimi K2.7 Code vs Claude Fable 5: Landing pages that cost 94% less
We ran an experiment where we had Kimi K2.7 Code and Claude Fable 5 each produce 12 landing pages for a side‑by‑side comparison. Overall, Kimi K2.7 Code cost about 94% less (16x less) than Fable 5 and...
1.8K
Together AI reposted
Vipul Ved Prakash
@vipulved
3h
Just added gobs of H100s, H200s and B200s on our on-demand compute platform.
NVIDIA GPU Clusters: H100, H200, B200, GB200 | Together AI
From together.ai
1.1K
Together AI reposted
Decagon
@DecagonAI
23h
🤝
Together AI
@togethercompute
Jun 16
Article
How Decagon Engineered Sub-Second Voice AI with Together AI
The challenge with running voice agents Voice latency is audible. Decagon’s leadership frames voice as the most demanding surface because latency is immediately perceptible. Long pauses create awkward...
2.2K
Together AI
@togethercompute
Jun 16
We tested closed and open models by asking them to build small, playable games. Open models were much cheaper and faster, while producing games that were often close in quality. → Opus 4.8 was 15x more expensive than MiniMax M3 → GPT-5.5 was 10x more expensive than Nemotron
Hassan
00:00
Hassan
6.8K
Together AI reposted
Hassan
@nutlope
Jun 16
Built a visual benchmark where I asked closed and open source models to build small games. Main takeaway: OSS models were a lot faster, cheaper, & produced games with similar quality. Specifically: * Opus 4.8 was 15x more expensive than MiniMax M3 * GPT-5.5 was 10x more
00:00
8.6K
Together AI
@togethercompute
Jun 16
.@DecagonAI cut voice agent cost per turn nearly 6x with Together AI. They moved from closed models to fine-tuned open models, while keeping latency low enough for real-time voice: → <400ms p95 model latency per turn → custom speculators and prompt caching → optimized
Together AI
@togethercompute
Jun 16
Article
How Decagon Engineered Sub-Second Voice AI with Together AI
The challenge with running voice agents Voice latency is audible. Decagon’s leadership frames voice as the most demanding surface because latency is immediately perceptible. Long pauses create awkward...
4.6K
Together AI
@togethercompute
Jun 16
Article
How Decagon Engineered Sub-Second Voice AI with Together AI
The challenge with running voice agents Voice latency is audible. Decagon’s leadership frames voice as the most demanding surface because latency is immediately perceptible. Long pauses create awkward...
6.6K
Together AI
@togethercompute
Jun 15
Optimizing GLM 5.1 came down to three things: -> Rewrote the indexer topk kernel -> Fused the indexer kernel to reduce memory and launch overhead -> Eliminated CPU overhead that was gating prefill throughput The bigger win was in the indexer. Once we fixed that, the rest made
15K
Together AI
@togethercompute
Jun 15
Try GLM 5.1 today:
GLM-5.1 API | Together AI
From together.ai
1.8K
Together AI
@togethercompute
Jun 15
Cartesia Sonic 3.5 is now available on Together AI. We added 150+ @cartesia Sonic 3.5 voices to voice finder, so developers can listen, compare, and pick the right voice for real-time agents before deploying on Together AI.
00:00
4.7K
Together AI
@togethercompute
Jun 15
Sonic 3.5 is built for real-time TTS where latency, pronunciation, and delivery matter. → Sub-90ms latency → Native support for 42 languages → Accurate transcript following for codes, IDs, and structured speech → Context-aware pronunciation for words like read, bass, and bow
2K
Together AI
@togethercompute
Jun 15
Try Sonic 3.5 in voice finder: voicefinder.together.ai Model page: together.ai/models/cartesi…
1.4K
Together AI reposted
Karan Goel
@krandiash
Jun 15
We released Sonic-3.5 and Ink-2, the #1 streaming models for text to speech and speech to text you can use in your voice agents today. New architectures enable new frontiers for speed and quality. We're now the only provider to have #1 models for both speaking and listening.
00:00
6.9M
Together AI
@togethercompute
Jun 14
DeepSeek V4 Pro on Together AI is now #1 on Artificial Analysis for both output speed and latency. Serving V4 well is an inference systems problem: KV cache, prefix reuse, kernels, and endpoint profiles. We break down the systems work here:
Together AI
@togethercompute
May 11
Article
Serving DeepSeek-V4: why million-token context is an inference systems problem
Benchmark tables miss the main point of DeepSeek-V4: the important change is architectural. V4 turns million-token context into a serving-systems problem. The model supports a 1M-token context window...
5.2K
Together AI reposted
Vipul Ved Prakash
@vipulved
Jun 14
DeepSeek V4 Pro on @togethercompute becomes #1 on both latency and speed.
9.2K
Together AI
@togethercompute
Jun 13
Kimi-K2.7-Code from @Kimi_Moonshot is now available on Together AI. Built on Kimi K2.6, it’s a coding-focused agentic model for long-horizon software engineering workflows, now running on Together’s research-powered inference stack for tool-heavy coding agents.
5.7K
Together AI
@togethercompute
Jun 13
Highlights: 👉 Stronger real-world coding performance over Kimi K2.6 👉 ~30% lower thinking-token usage for better token efficiency 👉 256K context for larger repos and longer coding sessions 👉 Interleaved thinking and multi-step tool calling for coding agents
1.7K
Together AI
@togethercompute
Jun 13
Try it now:
Kimi K2.7 Code API | Together AI
From together.ai
1.5K