specification: API Commons Plans
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/Plans
provider: Together AI
providerId: together-ai
created: '2026-05-08'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-08, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-08'
reconciled: true
tags:
- AI
- LLM
- Inference
- Open Source
- Fine-tuning
- GPU
- Plans
description: >-
  Together AI offers transparent, usage-based pricing across serverless inference,
  embeddings, rerank, image, video, and audio generation; managed fine-tuning;
  asynchronous batch with up to 50% discount; and hourly dedicated/reserved
  GPU compute. Per-model serverless rates vary widely - representative ranges
  are captured below; verify current rates on the pricing page.
notes: >-
  All token rates are per 1M tokens. Image/video/audio rates are per image,
  per video, or per 1M characters depending on modality. GPU rates are hourly.
sources:
- https://www.together.ai/pricing
- https://docs.together.ai/
plans:
- id: together-ai-serverless
  name: Serverless Inference (Pay-as-you-go)
  type: usage
  description: >-
    On-demand per-token / per-asset pricing for chat, embeddings, rerank,
    image, video, and audio across the Together model catalog. Start free
    with promotional credits, scale on demand without commitments.
  entries:
  - label: Chat (small)
    name: chat_small
    type: usage
    metric: tokens
    limit: -1
    timeFrame: month
    geo: global
    unit: 1000000
    price: from $0.10/$0.15 per 1M (e.g., Qwen3.5 9B input/output)
    userMultiplied: false
  - label: Chat (mid)
    name: chat_mid
    type: usage
    metric: tokens
    limit: -1
    timeFrame: month
    geo: global
    unit: 1000000
    price: ~$0.88/$0.88 per 1M (e.g., Llama 3.3 70B)
    userMultiplied: false
  - label: Chat (premium)
    name: chat_premium
    type: usage
    metric: tokens
    limit: -1
    timeFrame: month
    geo: global
    unit: 1000000
    price: $1.40-$7.00 per 1M (e.g., GLM-5.1, DeepSeek-R1)
    userMultiplied: false
  - label: Embeddings
    name: embeddings
    type: usage
    metric: tokens
    limit: -1
    timeFrame: month
    geo: global
    unit: 1000000
    price: from $0.02 per 1M (multilingual-e5-large-instruct)
    userMultiplied: false
  - label: Image Generation
    name: image_generation
    type: usage
    metric: images
    limit: -1
    timeFrame: month
    geo: global
    unit: 1
    price: $0.0006-$0.134 per image
    userMultiplied: false
  - label: Video Generation
    name: video_generation
    type: usage
    metric: videos
    limit: -1
    timeFrame: month
    geo: global
    unit: 1
    price: $0.14-$3.20 per video
    userMultiplied: false
  - label: Audio (TTS / STT)
    name: audio
    type: usage
    metric: characters
    limit: -1
    timeFrame: month
    geo: global
    unit: 1000000
    price: $0.0015-$65.00 per 1M characters
    userMultiplied: false
  elements:
  - name: Chat Completions
  - name: Embeddings
  - name: Rerank
  - name: Images
  - name: Video
  - name: Audio
  - name: Vision
- id: together-ai-batch
  name: Batch API
  type: usage
  description: >-
    Asynchronous batch inference at up to 50% discount over the equivalent
    serverless rate.
  entries:
  - label: Batch Tokens
    name: batch_tokens
    type: usage
    metric: tokens
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1000000
    price: up to 50% off serverless rates
    userMultiplied: false
  elements:
  - name: Batch Chat
  - name: Batch Embeddings
- id: together-ai-fine-tuning
  name: Fine-Tuning
  type: usage
  description: >-
    Supervised fine-tuning (LoRA and full) priced per 1M training tokens,
    scaled by base-model size.
  entries:
  - label: Up to 16B (LoRA / Full)
    name: ft_small
    type: usage
    metric: tokens
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1000000
    price: $0.48-$1.35 per 1M
    userMultiplied: false
  - label: 17B-69B
    name: ft_medium
    type: usage
    metric: tokens
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1000000
    price: $1.50-$4.12 per 1M
    userMultiplied: false
  - label: 70B-100B
    name: ft_large
    type: usage
    metric: tokens
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1000000
    price: $2.90-$8.00 per 1M
    userMultiplied: false
  - label: Specialized (DeepSeek-R1, GLM-5, Kimi)
    name: ft_specialized
    type: usage
    metric: tokens
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1000000
    price: $9-$100 per 1M
    userMultiplied: false
  elements:
  - name: Supervised Fine-Tuning
  - name: LoRA
  - name: Full Fine-Tuning
  - name: DPO
- id: together-ai-dedicated
  name: Dedicated Inference Endpoints
  type: usage
  description: >-
    Reserved single-tenant GPU-backed inference endpoints billed hourly.
  entries:
  - label: 1x H100 - 1x B200 (hourly)
    name: dedicated_hourly
    type: usage
    metric: hours
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1
    price: $3.99-$9.95 per hour
    userMultiplied: false
  elements:
  - name: H100
  - name: H200
  - name: B200
- id: together-ai-gpu-clusters
  name: GPU Clusters
  type: usage
  description: >-
    On-demand and reserved bare-metal GPU clusters for training and
    self-managed inference.
  entries:
  - label: On-demand (hourly)
    name: gpu_on_demand
    type: usage
    metric: hours
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1
    price: $3.49-$7.49 per hour
    userMultiplied: false
  - label: Reserved (7-30+ days)
    name: gpu_reserved
    type: usage
    metric: hours
    limit: -1
    timeFrame: usage
    geo: global
    unit: 1
    price: $2.99-$7.15 per hour (deeper discounts at 181+ days)
    userMultiplied: false
  elements:
  - name: H100 Cluster
  - name: H200 Cluster
  - name: B200 Cluster
- id: together-ai-enterprise
  name: Enterprise
  type: enterprise
  description: >-
    Volume commitments, custom capacity, dedicated regions, and
    procurement-friendly contracts. Contact Together AI sales.
  entries:
  - label: Enterprise Agreement
    name: enterprise
    type: flat
    metric: contract
    limit: -1
    timeFrame: year
    geo: global
    unit: 1
    price: contact sales
    userMultiplied: false
  elements:
  - name: Custom Volume Pricing
  - name: SLA
  - name: VPC
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com