← Back to Command Center
Model Availability & Pricing
Summary:
- Token pricing is broadly consistent across platforms when comparing equivalent models.
- Databricks pay-per-token pricing aligns closely with provider list pricing.
- Bedrock provides a ~4% discount for Anthropic (Claude) under our contract, making it the lowest-cost option for those models.
- Databricks exposes higher-end OpenAI and Google models (e.g., GPT-5, Gemini Pro variants) that are more capable, but also more expensive than the models currently surfaced via Bedrock.
Anthropic (Claude) — Per 1M Tokens
| Model |
Bedrock Input |
Bedrock Output |
Databricks Input |
Databricks Output |
| Opus 4.6 |
$4.80 |
$24.00 |
$5.00 |
$25.00 |
| Opus 4.5 |
$4.80 |
$24.00 |
$5.00 |
$25.00 |
| Opus 4.1 |
$14.40 |
$72.00 |
$15.00 |
$75.00 |
| Opus 4 |
$14.40 |
$72.00 |
$15.00 |
$75.00 |
| Sonnet 4.5 |
$2.88 |
$14.40 |
$3.00 |
$15.00 |
| Sonnet 4 |
$2.88 |
$14.40 |
$3.00 |
$15.00 |
| Sonnet 3.7 |
$2.88 |
$14.40 |
$3.00 |
$15.00 |
| Haiku 4.5 |
$0.96 |
$4.80 |
$1.00 |
$5.00 |
| Haiku 3.5 |
$0.768 |
$3.84 |
— |
— |
Bedrock prices reflect our 4% negotiated discount with Anthropic. Hitting Claude via direct API is the same cost as Databricks.
OpenAI — Per 1M Tokens
| Model |
Bedrock Input |
Bedrock Output |
Databricks Input |
Databricks Output |
| GPT OSS 20B |
$0.07 |
$0.30 |
— |
— |
| GPT OSS 120B |
$0.15 |
$0.60 |
— |
— |
| GPT OSS Safeguard 20B |
$0.07 |
$0.20 |
— |
— |
| GPT OSS Safeguard 120B |
$0.15 |
$0.60 |
— |
— |
| GPT-5.2 |
— |
— |
$1.75 |
$14.00 |
| GPT-5.1 |
— |
— |
$1.25 |
$10.00 |
| GPT-5.1 Codex Max |
— |
— |
$1.25 |
$10.00 |
| GPT-5 |
— |
— |
$1.25 |
$10.00 |
| GPT-5 mini |
— |
— |
$0.25 |
$2.00 |
| GPT-5.1 Codex Mini |
— |
— |
$0.25 |
$2.00 |
| GPT-5 nano |
— |
— |
$0.05 |
$0.40 |
OpenAI OSS models (available via Bedrock) can offer lower cost at scale, since you are paying primarily for compute rather than per-token API usage; however, this requires managing performance, reliability, and infrastructure. Databricks abstracts this complexity, handling deployment and scaling on your behalf.
Google — Per 1M Tokens
| Model |
Bedrock Input |
Bedrock Output |
Databricks Input |
Databricks Output |
| Gemini 3.0 / 3.1 Pro (Short) |
— |
— |
$2.50 |
$15.00 |
| Gemini 3.0 / 3.1 Pro (Long) |
— |
— |
$5.00 |
$22.50 |
| Gemini 3.0 Flash |
— |
— |
$0.63 |
$3.75 |
| Gemini 2.5 Pro (Short) |
— |
— |
$1.25 |
$10.00 |
| Gemini 2.5 Pro (Long) |
— |
— |
$2.50 |
$15.00 |
| Gemini 2.5 Flash |
— |
— |
$0.30 |
$2.50 |
| Gemma 3 4B |
$0.04 |
$0.08 |
— |
— |
| Gemma 3 12B |
$0.09 |
$0.29 |
— |
— |
| Gemma 3 27B |
$0.23 |
$0.38 |
— |
— |
Google Gemma models (available via Bedrock) can offer lower cost at scale, since you are paying primarily for compute rather than per-token API usage; however, this requires managing performance, reliability, and infrastructure. Databricks abstracts this complexity, handling deployment and scaling on your behalf.
← Back to Choosing an LLM Serving Path