Products10月29日 01:09

Together 的 post-training 服务采用按 GPU 计费，并行 rollout 越多单次成本越低

官方账号Together AI@togethercompute

Post

Why we picked it

Together 的 post-training 服务按 GPU 收费，并行跑的 rollout 越多单次越便宜，做 agent 训练烧钱的朋友可以算算账。

Together 官方账号发文说明其 post-training 服务的定价机制：按 GPU 计费，而非按 rollout 计费。在训练 agent 模型时，长 rollout 的开销会快速累积，但在同一 GPU 上并行运行多个 rollout 可以摊薄单个 rollout 的成本。对需要大量长 rollout 的 agent 强化学习训练场景，这种计费方式可以明显降低开销。

Original · Together AI

long agent rollouts can get expensive fast

together post-training uses per-gpu pricing, so running more rollouts in parallel brings the cost of each one down

https://t.co/0WHnc47hCY https://t.co/y7zHkdU6OD

#Together #post-training #智能体 #微调

Keep reading

Recent stories with the same tag or section

Products Builders01:08

Chinese contentTogether Post-Training 支持多 LoRA 实验共享同一 GPU 容量

Together 的 post-training 平台现在能让多个 LoRA 实验共用一块 GPU，想试更多微调想法不用再重复开环境，做 post-training 的可以看看。

#Together #LoRA #post-training #微调

官方账号Together AI@togethercomputeSource

Products Builders04:09

Chinese contentfactory 在 Beam 上线前提前测试 Reflection AI 模型

factory 提前拿到了 Together 上的 Beam，拿 Reflection AI 的模型跑真实编程任务，连工具报错恢复和仓库验证都测了，上线当天就已经摸透脾气。

#factory #Reflection AI #Together #Beam

官方账号Together AI@togethercomputeSource

Products Builders02:35

Chinese contentDeel 数据科学团队用 Together 后训练平台数小时内跑通 RL 实验

Deel 团队在 Together 的后训练平台几小时就跑起 RL 实验，一张 GPU 同时塞下 5 个实验，做强化学习训练的可以看看。

#Together #Deel #强化学习 #微调

官方账号Together AI@togethercomputeSource

Products Builders11:38

Chinese contentTogether 推出 post-training 服务，自定义训练全流程

Together 出了个 post-training 平台，数据和奖励函数都能自己写，LoRA 或全参数微调都行，GPU 底层他们包了，想自己调模型的话可以看看。

#Together #post-training #LoRA #微调

官方账号Together AI@togethercomputeSource

Products BuildersSelected01:55

Together cuts post-training price on B200 GPUs to $9 per hour

The discount makes B200-based SFT and RL runs cheaper for teams that want full control of their training loop rather than a managed pipeline. Startups fine-tuning open-source models should compare this against renting B200 capacity on clouds like AWS or CoreWeave before the eight-week window ends. Pricing is in US dollars and available to customers globally, including Southeast Asia.

#Together #B200 #fine-tuning #GPU pricing

官方账号Together AI@togethercomputeSource

Industry BuildersSelected73°01:11

Interconnects plot shows exponential growth

This visualization highlights the accelerating pace of AI infrastructure development, showing how quickly the market is expanding beyond major cloud providers. For developers and businesses in Southeast Asia, this growth signals increasing accessibility to competitive AI inference options beyond traditional providers like AWS or Google Cloud.

#Interconnects #Together #Baseten #Fireworks

Nathan Lambert@natolambertSource