DeepInfra vs Novita AI
SEPT 2026 auditA comparison of DeepInfra and Novita AI built from values read directly from each provider's own documentation, with the source recorded against every figure.
The short answer
Choose DeepInfra if…
Running open-weight models at the lowest published per-token rates here — DeepSeek-V4-Flash at $0.06 per million input is a fraction of what the lab itself charges — with dedicated H100s at $2.20 an hour if throughput demands it.
Editorial · Palash Bagchi · approved
Choose Novita AI if…
Breadth — 200+ models under one API with a cache-read rate published next to most of them, and an H100 SXM at $3.39 an hour if the same vendor should run the dedicated side too.
Editorial · Palash Bagchi · approved
Consider something else if…
- DeepInfra: Only some models carry per-token pricing; the rest bill for inference execution time, which is a different and harder thing to forecast.
- Novita AI: It is rarely the cheapest for any given model: DeepInfra serves DeepSeek-V4-Flash-0731 at $0.06 against Novita's $0.44, and its H100 is $2.20 against Novita's $3.39 on demand.
20 sourced criteria separate them — see where, with sources, below.
Pick the criteria you care about. The chart counts how many of them lean toward each provider — the same read as scanning the bars below, just totalled for the ones you chose.
DeepInfra 0
Novita AI 0
At a glance.
| Criterion | DeepInfra | Novita AI |
|---|---|---|
| Pricing | ||
| Free tier | No | Yes — Three inclusionAI Ling 3.0 Flash variants are listed at $0 input and $0 output; all other models are paid per token. |
| Kind of free offer | No free offer | Free tier |
| Entry paid plan | No minimum spend and no seat fees - billing is pay-per-token usage only, with no charge for idle GPU time. | No plan fee — credit is prepaid, and a manual top-up must be greater than $10. |
| Pricing model | Two products on one bill: per-token serverless inference with a separate cached-input rate for each model, and dedicated GPUs billed per minute and invoiced weekly. Models without per-token pricing are billed for inference execution time instead. | Per-token serverless inference with a cache-read rate published beside most models, and GPU instances offered at both on-demand and spot rates — the spot column is roughly half. Two models, Ling 3.0 Flash Fin and Sante, are listed at no charge on both input and output. |
| Infrastructure | ||
| Regions | Serverless inference runs on DeepInfra's own inference-optimised infrastructure in US-based data centers. No region selector is documented for the shared inference API. | GPU cloud page advertises 14 global regions in its feature list and, lower on the same page, 20+ locations across 4+ continents. |
| Inference | ||
| Input price, top model | $1.30 per million input tokens for DeepSeek-V4-Pro, the most expensive model in its catalogue. Cached input is $0.10. | $2.00 per million input tokens for Qwen3.8 Max, the most expensive model in its catalogue, with cache reads at $0.25 and output at $6.00. |
| Output price, top model | $2.60 per million output tokens for DeepSeek-V4-Pro. | $6.00 per million output tokens for Qwen3.8 Max. |
| Input price, cheapest model | $0.06 per million input tokens for DeepSeek-V4-Flash-0731, with cached input at $0.015 and output at $0.18. DeepSeek itself charges $0.44 at peak for the same checkpoint and Fireworks $0.22. | $0.075 per million input tokens for GLM 5.3 Flash at short context, rising to $0.15 above the threshold, with cache reads at $0.015 and output at $0.25. Fireworks serves the same model at $0.15. |
| Context window | 1024k tokens on the DeepSeek V4 models it serves; 160k on the V3 generation. The window is the model's rather than a platform limit. | 1M tokens on the DeepSeek V4 line and GLM 5.3, published as "1M" rather than an exact figure — so it is recorded at 1,000,000 and may in practice be the 1,048,576 other providers state. Qwen3.8 is listed at 977K. |
| Max output tokens | Hard cap of 16384 output tokens per response for most models on the chat completions API; the exact limit is model-dependent. | 128K (131,072) maximum output tokens for Qwen3.8 Max, the highest-priced model in the catalogue, against a 977K context. |
| Prompt caching | Supported | Supported |
| Batch discount | Batch API jobs are billed 20% below real-time pricing, with results returned within 24 hours. | Introductory 50% discount on input and output tokens for supported models when submitted through the Batch API. |
| Rate limits | Default 200 concurrent requests per model per account (429 above it); a limit increase can be requested in the dashboard. | Per-model RPM and TPM ceilings set by account tier T1–T5, where tier follows monthly top-ups: T1 up to $50 a month, T5 at least $10,000 in one of the last three months. |
| GPU | ||
| H100 SXM, per GPU-hour | $2.20 per GPU-hour for a dedicated H100 80GB — below every dedicated cloud in this dataset. A100 80GB is $0.89, H200 $2.69, B200 $3.69. Billed in minute granularity, invoiced weekly. | $3.39 per GPU-hour for an on-demand H100 SXM 80GB, or $1.70 spot. |
| Largest GPU offered | 270GB per GPU on a B300, at $4.89 per GPU-hour. | 80GB per GPU on an H100 SXM, the largest card it lists; below it are an L40S 48GB, RTX 5090 32GB and RTX 4090 24GB. |
| Compliance | ||
| SOC 2 | Security measures stated to comply with SOC 2 standards in the privacy policy; no report, audit type or auditor is published. | SOC 2 Type II; the audit report is available under NDA to enterprise customers on request through the Trust Center. |
| HIPAA | Technical and organizational measures for HIPAA compliance stated in the privacy policy; no BAA or eligible-services list is published. | Not by default configured to comply with HIPAA, FISMA or GLBA; the Terms of Service direct affected users to contact support for enterprise arrangements. Terms of Use section 1, page updated August 5, 2026. |
| GDPR / data residency | Privacy Policy: data may be transferred, stored and processed outside your jurisdiction, specifically in the United States; EU residents keep GDPR rights. | GDPR compliance claimed; EU/EEA transfers rely on Standard Contractual Clauses (Module 2), UK transfers on the IDTA or UK Addendum. |
| Encryption at rest | Inference inputs and outputs are not stored to disk; bulk-inference data may be stored on disk in encrypted form, then deleted. | AES-256 for data at rest; TLS 1.2+ in transit between customers, Novita and its service providers. |
| Role-based access control | Team roles are member and admin; an admin can change roles, but all team members share the same API tokens and models. | Team accounts define five roles — Owner, Admin, Developer, Basic and Billing — each with read/write or view-only rights over all or only self-created resources. |
| Dedicated infrastructure | Private model deployments run on dedicated GPU infrastructure - your weights, your endpoint, your isolation - not shared with other users. | Deployments give a dedicated inference endpoint on exclusive GPUs, unlike serverless endpoints that share compute with other users. |
Where they differ.
Free tier
- DeepInfra
- No
- Novita AI
- Yes — Three inclusionAI Ling 3.0 Flash variants are listed at $0 input and $0 output; all other models are paid per token.
Sources (2) →Sources ↓
- Simple Pricing | Machine Learning Infrastructure | DeepInfra ↗
“You have to add a card or pre-pay [...] Pre-paid accounts draw down their balance as they go, and invoiced accounts are billed once at the beginning of the following month.”
Read 2026-09-13 · official pricing · Report a correction
- Pricing | Novita AI ↗
“Ling 3.0 Flash Fin 256K Free Free [...] Ling 3.0 Flash VL 256K Free Free [...] Ling 3.0 Flash Sante 256K Free Free”
Read 2026-09-13 · official pricing · Report a correction
Kind of free offer
- DeepInfra
- No free offer
- Novita AI
- Free tier
Sources (2) →Sources ↓
- Simple Pricing | Machine Learning Infrastructure | DeepInfra ↗
“You have to add a card or pre-pay [...] Pre-paid accounts draw down their balance as they go, and invoiced accounts are billed once at the beginning of the following month.”
Read 2026-09-13 · official pricing · Report a correction
- Pricing | Novita AI ↗
“Ling 3.0 Flash Fin 256K Free Free [...] Ling 3.0 Flash VL 256K Free Free [...] Ling 3.0 Flash Sante 256K Free Free”
Read 2026-09-13 · official pricing · Report a correction
Entry paid plan
- DeepInfra
- No minimum spend and no seat fees - billing is pay-per-token usage only, with no charge for idle GPU time.
- Novita AI
- No plan fee — credit is prepaid, and a manual top-up must be greater than $10.
Sources (2) →Sources ↓
- What is DeepInfra - DeepInfra ↗
“You only pay per token - no idle GPU time, no minimums, no seat fees.”
Read 2026-09-13 · official docs · Report a correction
- Payment Methods — Novita AI Documentation ↗
“You can manually top up a specified amount of credit, which must be greater than $10.”
Read 2026-09-13 · official docs · Report a correction
Pricing model
- DeepInfra
- Two products on one bill: per-token serverless inference with a separate cached-input rate for each model, and dedicated GPUs billed per minute and invoiced weekly. Models without per-token pricing are billed for inference execution time instead.
- Novita AI
- Per-token serverless inference with a cache-read rate published beside most models, and GPU instances offered at both on-demand and spot rates — the spot column is roughly half. Two models, Ling 3.0 Flash Fin and Sante, are listed at no charge on both input and output.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“Some of our langauge models offer per token pricing. Most other models are billed for inference execution time. [...] Billed in minute granularity Invoiced weekly”
Read 2026-09-06 · official pricing · Report a correction
- Pricing — Novita AI ↗
“Ling 3.0 Flash Fin 256K Free Free [...] Ling 3.0 Flash Sante 256K Free Free [...] On-Demand SPOT”
Read 2026-09-07 · official pricing · Report a correction
Regions
- DeepInfra
- Serverless inference runs on DeepInfra's own inference-optimised infrastructure in US-based data centers. No region selector is documented for the shared inference API.
- Novita AI
- GPU cloud page advertises 14 global regions in its feature list and, lower on the same page, 20+ locations across 4+ continents.
Sources (2) →Sources ↓
- Machine Learning Models and Infrastructure | DeepInfra ↗
“DeepInfra runs on our own cutting-edge inference optimised infrastructure, in secure US-based data centers.”
Read 2026-09-13 · official site · Report a correction
- GPU Cloud - High-Performance GPUs for AI | Novita AI ↗
“14 global regions Low-latency backbone with private peering. [...] 20+ locations 4+ continents”
Read 2026-09-13 · official site · Report a correction
Input price, top model
- DeepInfra
- $1.30 per million input tokens for DeepSeek-V4-Pro, the most expensive model in its catalogue. Cached input is $0.10.
- Novita AI
- $2.00 per million input tokens for Qwen3.8 Max, the most expensive model in its catalogue, with cache reads at $0.25 and output at $6.00.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“DeepSeek-V4-Pro 1024k $1.30 / $0.10 cached $2.60”
Read 2026-09-06 · official pricing · Report a correction
- Pricing — Novita AI ↗
“Qwen3.8 Max 977K $2 /Mt · Cache Read $0.25 /Mt $6 /Mt”
Read 2026-09-07 · official pricing · Report a correction
Output price, top model
- DeepInfra
- $2.60 per million output tokens for DeepSeek-V4-Pro.
- Novita AI
- $6.00 per million output tokens for Qwen3.8 Max.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“DeepSeek-V4-Pro 1024k $1.30 / $0.10 cached $2.60”
Read 2026-09-06 · official pricing · Report a correction
- Pricing — Novita AI ↗
“Qwen3.8 Max 977K $2 /Mt · Cache Read $0.25 /Mt $6 /Mt”
Read 2026-09-07 · official pricing · Report a correction
Input price, cheapest model
- DeepInfra
- $0.06 per million input tokens for DeepSeek-V4-Flash-0731, with cached input at $0.015 and output at $0.18. DeepSeek itself charges $0.44 at peak for the same checkpoint and Fireworks $0.22.
- Novita AI
- $0.075 per million input tokens for GLM 5.3 Flash at short context, rising to $0.15 above the threshold, with cache reads at $0.015 and output at $0.25. Fireworks serves the same model at $0.15.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“DeepSeek-V4-Flash-0731 1024k $0.06 / $0.015 cached $0.18”
Read 2026-09-06 · official pricing · Report a correction
- Pricing — Novita AI ↗
“GLM 5.3 Flash 1M $0.075 /Mt · Cache Read $0.015 /Mt $0.15 /Mt · Cache Read $0.03 /Mt $0.25 /Mt $0.5 /Mt”
Read 2026-09-07 · official pricing · Report a correction
Context window
1024k tokens on the DeepSeek V4 models it serves; 160k on the V3 generation. The window is the model's rather than a platform limit.
1M tokens on the DeepSeek V4 line and GLM 5.3, published as "1M" rather than an exact figure — so it is recorded at 1,000,000 and may in practice be the 1,048,576 other providers state. Qwen3.8 is listed at 977K.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“DeepSeek-V4-Pro 1024k [...] DeepSeek-V3.2 160k $0.26 / $0.13 cached $0.38”
Read 2026-09-06 · official pricing · Report a correction
- Pricing — Novita AI ↗
“DeepSeek V4 Pro 0813 1M [...] GLM 5.3 1M [...] Qwen3.8 Max 977K”
Read 2026-09-07 · official pricing · Report a correction
Max output tokens
- DeepInfra
- Hard cap of 16384 output tokens per response for most models on the chat completions API; the exact limit is model-dependent.
- Novita AI
- 128K (131,072) maximum output tokens for Qwen3.8 Max, the highest-priced model in the catalogue, against a 977K context.
Sources (2) →Sources ↓
- Chat Completions - DeepInfra ↗
“The maximum number of tokens that can be generated in a single response is model-dependent, with a hard cap of 16384 tokens for most models.”
Read 2026-09-13 · official docs · Report a correction
- Qwen3.8 Max API & Playground | Novita AI ↗
“Context Length 977K Max Output 128K”
Read 2026-09-13 · official site · Report a correction
Batch discount
- DeepInfra
- Batch API jobs are billed 20% below real-time pricing, with results returned within 24 hours.
- Novita AI
- Introductory 50% discount on input and output tokens for supported models when submitted through the Batch API.
Sources (2) →Sources ↓
- Introduction to Batch API - DeepInfra ↗
“get the results back within 24 hours, billed at 20% below real-time pricing”
Read 2026-09-13 · official docs · Report a correction
- Pricing | Novita AI ↗
“Batch inference is available at an introductory 50% discount on input and output tokens for supported models.”
Read 2026-09-13 · official pricing · Report a correction
Rate limits
- DeepInfra
- Default 200 concurrent requests per model per account (429 above it); a limit increase can be requested in the dashboard.
- Novita AI
- Per-model RPM and TPM ceilings set by account tier T1–T5, where tier follows monthly top-ups: T1 up to $50 a month, T5 at least $10,000 in one of the last three months.
Sources (2) →Sources ↓
- Rate Limits - DeepInfra ↗
“Every account has a default limit of 200 concurrent requests per model [...] You can request a rate limit increase in your”
Read 2026-09-13 · official docs · Report a correction
- Rate limits — Novita AI Documentation ↗
“Each account has a default rate limit for model calls, measured in RPM (requests per model per minute) and TPM (tokens per model per minute). [...] T1 Monthly top-ups did not exceed $50 in any of the last 3 calendar months. [...] T5 Monthly top-ups were at least $10,000 in at least one of the last 3 calendar months.”
Read 2026-09-13 · official docs · Report a correction
H100 SXM, per GPU-hour
- DeepInfra
- $2.20 per GPU-hour for a dedicated H100 80GB — below every dedicated cloud in this dataset. A100 80GB is $0.89, H200 $2.69, B200 $3.69. Billed in minute granularity, invoiced weekly.
- Novita AI
- $3.39 per GPU-hour for an on-demand H100 SXM 80GB, or $1.70 spot.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“A100 80GB $0.89 / GPU-hour H100 80GB $2.20 / GPU-hour H200 141GB $2.69 / GPU-hour B200 180GB $3.69 / GPU-hour B300 270GB $4.89 / GPU-hour [...] Billed in minute granularity Invoiced weekly”
Read 2026-09-06 · official pricing · Report a correction
- Novita AI GPU Cloud ↗
“On-Demand SPOT H100 SXM 80GB 80 GB VRAM 3.39/hr/GPU 1.70/hr/GPU RTX 4090 24GB 24 GB VRAM 0.33/hr/GPU 0.17/hr/GPU”
Read 2026-09-14 · official pricing · Report a correction
Largest GPU offered
270GB per GPU on a B300, at $4.89 per GPU-hour.
80GB per GPU on an H100 SXM, the largest card it lists; below it are an L40S 48GB, RTX 5090 32GB and RTX 4090 24GB.
Sources (2) →Sources ↓
- Pricing — DeepInfra ↗
“B300 270GB $4.89 / GPU-hour”
Read 2026-09-06 · official pricing · Report a correction
- GPU Instance — Novita AI ↗
“H100 SXM 80GB 80 GB VRAM 3.39/hr/GPU 1.70/hr/GPU [...] NVIDIA L40S 48GB 48 GB VRAM 0.55/hr/GPU”
Read 2026-09-07 · official pricing · Report a correction
SOC 2
- DeepInfra
- Security measures stated to comply with SOC 2 standards in the privacy policy; no report, audit type or auditor is published.
- Novita AI
- SOC 2 Type II; the audit report is available under NDA to enterprise customers on request through the Trust Center.
Sources (2) →Sources ↓
- Privacy Policy — DeepInfra ↗
“Our security measures comply with SOC 2 and ISO 27001 standards and include technical and organizational measures for GDPR and HIPAA”
Read 2026-09-13 · official site · Report a correction
- Novita AI Trust Center — FAQ ↗
“SOC 2 Type II: Available under NDA for enterprise customers [...] Detailed audit reports are available under NDA to enterprise customers upon request.”
Read 2026-09-13 · official site · Report a correction
HIPAA
- DeepInfra
- Technical and organizational measures for HIPAA compliance stated in the privacy policy; no BAA or eligible-services list is published.
- Novita AI
- Not by default configured to comply with HIPAA, FISMA or GLBA; the Terms of Service direct affected users to contact support for enterprise arrangements. Terms of Use section 1, page updated August 5, 2026.
Sources (2) →Sources ↓
- Privacy Policy — DeepInfra ↗
“Our security measures comply with SOC 2 and ISO 27001 standards and include technical and organizational measures for GDPR and HIPAA”
Read 2026-09-13 · official site · Report a correction
- Novita AI Terms of Service ↗
“The Site is not by default configured to comply with certain industry-specific regulations (including HIPAA, FISMA, and GLBA). If your use of the Site or Marketplace Offerings is subject to such laws, please contact us at support@novita.ai to discuss enterprise compliance arrangements before proceeding.”
Read 2026-09-13 · official site · Report a correction
GDPR / data residency
- DeepInfra
- Privacy Policy: data may be transferred, stored and processed outside your jurisdiction, specifically in the United States; EU residents keep GDPR rights.
- Novita AI
- GDPR compliance claimed; EU/EEA transfers rely on Standard Contractual Clauses (Module 2), UK transfers on the IDTA or UK Addendum.
Sources (2) →Sources ↓
- DeepInfra Privacy Policy ↗
“Your data may be transferred, stored, and processed outside of your jurisdiction, specifically in the United States. [...] If you are a resident of the European Union you may have additional rights under the General Data Protection Regulation.”
Read 2026-09-13 · official site · Report a correction
- Novita AI Trust Center — FAQ ↗
“GDPR Compliance: Aligned with EU requirements [...] Standard Contractual Clauses (SCCs) — European Commission Decision 2021/914 (Module 2: Controller-to-Processor), as appropriate to your deployment model”
Read 2026-09-13 · official site · Report a correction
Encryption at rest
- DeepInfra
- Inference inputs and outputs are not stored to disk; bulk-inference data may be stored on disk in encrypted form, then deleted.
- Novita AI
- AES-256 for data at rest; TLS 1.2+ in transit between customers, Novita and its service providers.
Sources (2) →Sources ↓
- Data Privacy - DeepInfra ↗
“Input data is not stored to disk during inference [...] we may need to store data for a longer period, potentially on disk in encrypted form”
Read 2026-09-13 · official docs · Report a correction
- Novita AI Trust Center — FAQ ↗
“Data at rest: AES-256 encryption for stored data [...] Data in transit: TLS 1.2+ encryption between you and our services, and between us and our service providers”
Read 2026-09-13 · official site · Report a correction
Role-based access control
- DeepInfra
- Team roles are member and admin; an admin can change roles, but all team members share the same API tokens and models.
- Novita AI
- Team accounts define five roles — Owner, Admin, Developer, Basic and Billing — each with read/write or view-only rights over all or only self-created resources.
Sources (2) →Sources ↓
- Okta SSO - DeepInfra ↗
“Admin can change team member roles (member / admin) [...] All team members share the same API tokens and models [...] For per-user isolation (separate tokens and models per person), contact us”
Read 2026-09-13 · official docs · Report a correction
- Team — Novita AI Documentation ↗
“Role is a set of permissions that define what actions a member can perform within a team. There are 5 roles in one team: [...] Action : Either Read/Write (create, update, delete, query) or View (query and list only). [...] Owner The highest level of permission. Manages the team, billing, and all resources. Each team has exactly one Owner.”
Read 2026-09-13 · official docs · Report a correction
Dedicated infrastructure
- DeepInfra
- Private model deployments run on dedicated GPU infrastructure - your weights, your endpoint, your isolation - not shared with other users.
- Novita AI
- Deployments give a dedicated inference endpoint on exclusive GPUs, unlike serverless endpoints that share compute with other users.
Sources (2) →Sources ↓
- Deploy Private Models - DeepInfra ↗
“DeepInfra allows you to deploy your own models on dedicated infrastructure [...] data stays on dedicated infrastructure, not shared with other users”
Read 2026-09-13 · official docs · Report a correction
- Novita Deployments User Guide — Novita AI Documentation ↗
“Unlike Serverless endpoints that share compute resources with other users, each Deployment gives you: [...] All compute resources are yours alone — no noisy neighbors”
Read 2026-09-13 · official docs · Report a correction
When these numbers change, hear about it. Sources are re-checked monthly; a repricing goes out as a short note.
Documented by only one.
These criteria are published by one provider and not the other. An absence here means we have not found a source, not that the feature is missing — except where we say we read the other provider's own page on a given date and it does not state it. Those rows lean, and the provider can correct them with a link.
Questions this comparison answers.
Entry paid plan: DeepInfra or Novita AI?
DeepInfra: No minimum spend and no seat fees - billing is pay-per-token usage only, with no charge for idle GPU time.
Novita AI: No plan fee — credit is prepaid, and a manual top-up must be greater than $10.
Free tier: DeepInfra or Novita AI?
DeepInfra: No
Novita AI: Yes — Three inclusionAI Ling 3.0 Flash variants are listed at $0 input and $0 output; all other models are paid per token.
Should I choose DeepInfra or Novita AI?
Pick DeepInfra if Running open-weight models at the lowest published per-token rates here — DeepSeek-V4-Flash at $0.06 per million input is a fraction of what the lab itself charges — with dedicated H100s at $2.20 an hour if throughput demands it.
Pick Novita AI if Breadth — 200+ models under one API with a cache-read rate published next to most of them, and an H100 SXM at $3.39 an hour if the same vendor should run the dedicated side too.
Comments are hosted by EchoThread, which sets cookies and offers social sign-in. Like the analytics on this site, it does not load until you accept — see the privacy page.