IBM watsonx.ai
- Categories
- Hyperscaler AI platforms
- Owned by
- IBM (NYSE: IBM) · also owns IBM Cloud
- HQ
- US
- Website
- https://www.ibm.com/products/watsonx-ai
- Pricing page
- https://www.ibm.com/products/watsonx-ai/pricing
- Fee model
- per_token+per_second_gpu
- Fees
- Pay-as-you-go inference billed in Resource Units (1 RU = 1,000 tokens); prices shown per 1M tokens. Most catalogue models are "Deploy on demand only" = hourly GPU hosting (e.g. 1xH100 $14.50/h, 8xH100 $116/h, 1xH200 $16/h, 8xH200 $128/h). Mistral commercial models add a model access fee.
- Free tier
- Free trial available
- OpenAI-compatible API
- partial
- Bring your own key
- no
- Notes
- IBM labels prices "indicative, may vary by country". Only a handful of models have pay-as-you-go token pricing.
- Sources
- https://www.ibm.com/products/watsonx-ai/pricing https://www.ibm.com/products/watsonx-ai/foundation-models collected 2026-09-18
LLM prices (7 models, USD per 1M tokens)
| Model | Input | Output | Cached input | Context | Source |
|---|---|---|---|---|---|
| gpt-oss-120b gpt-oss-120b |
$0.159 | $0.636 | — | — | ✓ src src2 |
| granite-4h-small granite-4h-small |
$0.064 | $0.265 | — | — | ✓ src src2 |
| Llama 3.3 70B Instruct llama-3-3-70b-instruct |
$0.753 | $0.753 | — | — | ✓ src src2 |
| Llama 4 Maverick 17B Instruct llama-4-maverick-17b-128e-instruct-fp8 |
$0.371 | $1.48 | — | — | ✓ src src2 |
| Mistral Large 3 (675B) mistral-large-2512 |
$0.636 | $1.91 | — | — | ✓ src src2 |
| Mistral Medium 3 mistral-medium-2505 |
$3.18 | $9.50 | — | — | ✓ src src2 |
| mistral-small-3.1-24b-instruct mistral-small-3-1-24b-instruct-2503 |
$0.106 | $0.318 | — | — | ✓ src src2 |