AI Tools
Together AI
Together AI : prix, capacités, mises à jour et alternatives, lus sur les pages de l’éditeur.

En anglais
Key facts
- Price
- Paid plans: see pricing Read on together.ai
- Free plan
- No Read on together.ai
- Platforms
- Web · API · Linux
- API
- Yes Read on together.ai
- Open source
- Not stated on the maker’s pages
- Last update
- 16 Nov 2026 Read on together.ai
- Prices checked
- 26 Oct 2026 Read on together.ai
Together AI is a developer platform for serverless and dedicated model inference, batch processing, fine-tuning, GPU clusters, sandboxes, and managed storage. It supports text, image, audio, video, transcription, embeddings, reranking, and moderation workloads.
Developers and AI teams use it to build applications, agents, media features, and production model deployments. Pricing is usage-based rather than subscription-based; dedicated infrastructure, compute, storage, and fine-tuning cost extra, and the platform requires at least $5 in purchased credits.
- Category
- Developer Tools
- Read from the maker’s pages
- 2026-09-19
Company
Features
- Run serverless inference for text, image, audio, video, transcription, and embedding models
- Process large workloads asynchronously with Batch Inference
- Deploy models on dedicated, single-tenant GPU infrastructure
- Fine-tune open models with supervised fine-tuning and preference optimization
- Run GPU clusters for training and inference workloads
- Create secure code sandboxes and code interpreter sessions through an API
- Use managed storage with parallel filesystems and no egress fees
Use cases
- Build production AI applications with model APIs
- Deploy coding, research, and customer-support agents
- Fine-tune open models for domain-specific behavior
- Process batch text, image, audio, or video workloads
- Run training and inference jobs on managed GPU clusters
- Add voice, media generation, or multimodal features to products
Pricing
Paid plans: see pricing
- Prices checked
- 2026-10-26
Serverless Inference
$0.0015–$4.50 per 1M tokens
- $0.0015–$4.50 per 1M tokens
- $0.0006–$0.134 per image
- $0.14–$3.20 per video
- $0.0015 per audio minute
- High-performance inference as APIs
- Chat, vision, embeddings, rerank
- Image and video generation
- Audio transcription
Dedicated Inference — NVIDIA HGX H100
$5.49 per GPU per hour
- $5.49 per GPU per hour
- Single-tenant GPU instances
- Guaranteed performance
- Support for custom models
- Autoscaling & traffic spike handling
Dedicated Inference — NVIDIA HGX H200
price on request
- Single-tenant GPU instances
- Guaranteed performance
- Support for custom models
Dedicated Inference — NVIDIA HGX B200
$8.99 per GPU per hour
- $8.99 per GPU per hour
- Single-tenant GPU instances
- Guaranteed performance
- Support for custom models
- Autoscaling & traffic spike handling
Dedicated Inference — NVIDIA HGX B300
price on request
- Single-tenant GPU instances
- Guaranteed performance
- Support for custom models
Dedicated Inference — NVIDIA GB200 NVL72
price on request
- Single-tenant GPU instances
- Guaranteed performance
- Support for custom models
Dedicated Inference — NVIDIA GB300 NVL72
price on request
- Single-tenant GPU instances
- Guaranteed performance
- Support for custom models
GPU Clusters — NVIDIA HGX H100
$1.99 per GPU per hour (Preemptible Compute)
- $1.99 per GPU per hour (Preemptible Compute)
- $3.99 per GPU per hour (ON-Demand)
- $3.69 per GPU per hour (Reserved, 3–30 days)
- $3.45 per GPU per hour (Reserved, 31–90 days)
- $3.19 per GPU per hour (Reserved, 91–180 days)
- Reliable GPU clusters at scale
- On-demand, preemptible and reserved capacity
GPU Clusters — NVIDIA HGX H200
$2.99 per GPU per hour (Preemptible Compute)
- $2.99 per GPU per hour (Preemptible Compute)
- $5.99 per GPU per hour (ON-Demand)
- $4.99 per GPU per hour (Reserved, 3–30 days)
- $4.15 per GPU per hour (Reserved, 31–90 days)
- $3.99 per GPU per hour (Reserved, 91–180 days)
- Reliable GPU clusters at scale
- On-demand, preemptible and reserved capacity
GPU Clusters — NVIDIA HGX B200
$4.09 per GPU per hour (Preemptible Compute)
- $4.09 per GPU per hour (Preemptible Compute)
- $8.19 per GPU per hour (ON-Demand)
- $7.99 per GPU per hour (Reserved, 3–30 days)
- $7.79 per GPU per hour (Reserved, 31–90 days)
- $6.79 per GPU per hour (Reserved, 91–180 days)
- Reliable GPU clusters at scale
- On-demand, preemptible and reserved capacity
Capabilities
Latest updates
- Last year we built megakernels. This year we deleted them: The road to Together Compiler
Together Compiler uses AI agents to rewrite and fuse GPU kernels, reaching 2,118 tokens/s on Llama 3.2 1B on B200.
- How we make Together AI easy for coding agents to use
Updates to docs, skills, and API errors raised coding-agent task completion from 75% to 97%.
- Together Compiler: Continual Compilation for AI inference
Together Compiler uses agents and persistent engines to optimize open models in under 24 hours, with up to 3.7x higher throughput as traffic changes.
- Polymorphic Inference: self-tuning inference for open models
Polymorphic Inference adapts to changing models, traffic and chips; Together Compiler uses AI agents to keep production inference tuned.
- Auto repair for Together GPU Clusters: from GPU fault to healthy node
Together GPU Clusters auto repair detects GPU, host, and fabric faults and lets you choose which repairs run automatically.
Where it works
Web · API · Linux
Alternatives to Together AI
- Baseten
Baseten provides infrastructure to deploy and run AI models, train models, and build agent workflows using managed inference, APIs, training tools, and sandboxes.
- Modal
Modal runs containerized applications on serverless cloud infrastructure, including GPU inference, training, batch jobs, and isolated code sandboxes.
- Fireworks AI
Fireworks AI hosts, routes, and trains open and custom models for applications, teams, and research.
- Cerebras Inference
Cerebras Inference serves AI models through a cloud API and web playground, with pay-per-token access for developers and reserved options for enterprises.
- Lambda 1-Click Clusters
Reserved, interconnected GPU clusters for training, fine-tuning, and inference on large AI workloads.
- Liquid Foundation Models
Liquid Foundation Models are open-weight AI models for text, vision, audio, and decision tasks, usable locally or through Liquid's API.
Questions about Together AI
Is Together AI free?
No. The maker’s pricing page lists no free plan for Together AI. Prices checked on 26 Oct 2026.
How much does Together AI cost?
The plans on the maker’s pricing page: Serverless Inference ($0.0015–$4.50 per 1M tokens), Dedicated Inference — NVIDIA HGX H100 ($5.49 per GPU per hour), Dedicated Inference — NVIDIA HGX H200 (price on request), Dedicated Inference — NVIDIA HGX B200 ($8.99 per GPU per hour), Dedicated Inference — NVIDIA HGX B300 (price on request), Dedicated Inference — NVIDIA GB200 NVL72 (price on request), Dedicated Inference — NVIDIA GB300 NVL72 (price on request), GPU Clusters — NVIDIA HGX H100 ($1.99 per GPU per hour (Preemptible Compute)), GPU Clusters — NVIDIA HGX H200 ($2.99 per GPU per hour (Preemptible Compute)) and GPU Clusters — NVIDIA HGX B200 ($4.09 per GPU per hour (Preemptible Compute)). Prices checked on 26 Oct 2026.
Does Together AI have an API?
Yes. The maker’s pages state that Together AI has an API. Read on the maker’s pages on 19 Sept 2026.
Which platforms does Together AI run on?
Together AI runs on Web, API and Linux, as its maker’s pages state. Read on the maker’s pages on 19 Sept 2026.
Is Together AI still available?
Yes, as of 19 Sept 2026: its maker’s pages were read that day and carried no notice that Together AI has ended.
What changed recently in Together AI?
The latest entry in its maker’s changelog: 16 Nov 2026, “Last year we built megakernels. This year we deleted them: The road to Together Compiler”. Before it: 16 Nov 2026, “How we make Together AI easy for coding agents to use”. Changelog checked on 18 Nov 2026.
Sources
- https://www.together.ai/pricing — Official pricing page with serverless inference, dedicated inference, GPU cluster, sandbox, storage, and fine-tuning prices.
- https://www.together.ai/products — Official product page describing inference, compute, model shaping, sandbox, storage, and evaluation products.
- https://www.together.ai/blog/together-ai-expands-fine-tuning-service-with-more-models-live-metrics-and-finer-controls — Official September 11, 2026 product update covering fine-tuning features and selected price reductions.
- https://support.together.ai/articles/1862638756-changes-to-free-tier-and-billing-july-2025 — Official support article stating that Together AI does not offer free trials and requires a minimum $5 credit purchase.
- https://www.together.ai/terms-of-service — Official terms identifying Together Computer, Inc. and describing programmatic APIs and web interfaces.
Not confirmed
- Latest version
- Together AI does not print a current product version or proprietary model name on the reviewed official pages.
- Open source
- The reviewed official pages describe open-model support but do not state that the Together AI platform itself is open source.
- Platforms
- Linux was inferred from GPU cluster and compute documentation context; no complete official platform list was found.