Platform comparison
Novita AI vs DeepInfra: costs, performance and limits
Choose DeepInfra for the lower GPT-OSS 120B request estimate ($0.00054 vs $0.00075), or Novita AI for about 6.3× the Artificial Analysis (AA) output rate (228 vs 36 tokens/s).
| Dimension | Winner | Margin | Evidence |
|---|---|---|---|
| Estimated request cost | DeepInfra | $0.00021 difference per example | DeepInfra $0.00054 vs Novita AI $0.0007523456 |
| Output speed (tokens/s) | Novita AI | 6.3× output rate | Novita AI 228 vs DeepInfra 36 tokens/s1 |
| Input / 1M tokens | DeepInfra | $0.013 unit-price gap | DeepInfra $0.037 vs Novita AI $0.05 / 1M tokens23456 |
Performance
Performance
Swipe to see all columns
Pricing
Pricing
Swipe to see all columns
Capabilities and serving conditions
Capabilities and serving conditions
| Dimension | Novita AI | DeepInfra |
|---|---|---|
| API and model ID | https://api.novita.ai/openaiopenai/gpt-oss-120b234 | https://api.deepinfra.com/v1/openaiopenai/gpt-oss-120b56 |
| Model catalog | 200+ | 389 entries in the public catalog snapshot; catalog scopes differ from Novita’s advertised 200+ model count.7 |
| Billing basis | Published input and output token rates; dedicated capacity and discounts are separate. | Published input and output token rates; dedicated capacity and discounts are separate. |
| Deployment options | Serverless API and dedicated endpoints. | Serverless API and private endpoints. |
Which to choose
When each option fits
Novita AI
- Prefer Novita AI for long streamed replies: AA records 228 tokens/s for Novita AI vs 36 tokens/s on DeepInfra's standard endpoint.
- Its $0.00075 request estimate is $0.00021 above DeepInfra; this choice favors output rate over the lowest token cost.
- Keep text, image, video and audio APIs alongside Agent Sandbox on one Novita account.
DeepInfra
- Prefer DeepInfra for uncached token budgets: the request estimate is $0.00054 versus Novita AI’s $0.00075.
- Its first-chunk median is 0.71 s versus 0.91 s, while sustained output is slower at 36 versus 228 tokens/s.
- Use the standard endpoint for the prices shown here; DeepInfra Turbo has separate pricing and performance.
Methodology
Methodology and sources
The GPT-OSS 120B comparison uses Novita AI's and DeepInfra's standard endpoints in AA's 2026-10-15 provider table: high reasoning effort, 10,000 input tokens and 72-hour medians. DeepInfra Turbo is a separate endpoint and is excluded. The first chunk can contain reasoning; Endpoint Accuracy Index v1.0 is relative to a reference endpoint, not a task pass rate. DeepInfra's counted catalog entries and Novita's advertised model total have different inventory scopes.
DeepInfra's standard model page and Novita's GPT-OSS 120B page supply the uncached synchronous token rates. The chart estimates 10,000 input and 1,000 billed output tokens, including charged reasoning, before tax, retries, discounts and dedicated-capacity charges. It is not an invoice or a comparison of Turbo pricing. Novita AI publishes the page using public sources; no private endpoint benchmark was run.
- Artificial Analysis · GPT-OSS 120B Providers artificialanalysis.ai · Checked 2026-10-15
- Novita AI · GPT-OSS 120B novita.ai · Checked 2026-10-15
- Novita AI · Platform novita.ai · Checked 2026-10-15
- Novita AI · Batch docs.novita.ai · Checked 2026-10-15
- DeepInfra · GPT-OSS 120B deepinfra.com · Checked 2026-10-15
- DeepInfra · Rate Limits deepinfra.com · Checked 2026-10-15
- DeepInfra · Model Catalogapi.deepinfra.com · Checked 2026-10-15
Build with Novita Model APIs
Create an API key and connect your application.