SOVINFRAPrivate modeless inference · EU-hosted.

European AI inference

Inference built around your workload.

Fast, cost-efficient inference for AI agents, voice AI and business automation. Choose your model or deploy dedicated European capacity at serverless pricing. Pay per token.

Your first billion tokens are free.

No credit card required.

14-day trial · 100 million tokens per day · Up to 2 requests processed concurrently per account · Up to 160,000 input tokens per Qwen or Gemma request.

Get API key · 1B free tokens

Status

First-token target under 120 ms, excluding queue time · Redundant hosting designed for 99.95% availability

Service status →

Built for production AI

One URL. The right model. One meter.

A unified inference layer for direct models, SovAPI routing and specialized workloads — served on European infrastructure operated by SovInfra.
01

Migrate in one line

Point any OpenAI-compatible client to SovInfra. Call a model directly or use SovAPI routing without a custom integration.

02

Operated & specialized models

Gemma 4 31B, Qwen 3.8-27B, Whisper, BGE-M3 and Kokoro are tested, monitored and served by SovInfra. Specialized open-weight models can be deployed for your workload.

03

Test it live →

Run your own prompt on SovInfra and see TTFT and generation speed without creating an account.

04

European by design

No intermediary, no data retention and no training on customer data. Built for repetitive, context-heavy agent loops.

Dedicated. Pay per token.

Priority and high-performance capacity built around your workload. OpenAI-compatible API.

Need dedicated capacity or a specialized model?

How it works

  1. Tell us about your workload.

    Answer five questions about your use case and production profile.

  2. Test on your real traffic.

    Use your free token allowance to measure performance on your own workload.

  3. We configure the engine.

    We size the deployment around your model, cache usage, context length, and concurrent requests. We provide a written proposal specifying the GPU capacity reserved for your environment.

Tell us about your workload →

Automatic routing

Let SovAPI choose
"model": "sovapi"
Automatic routing across Gemma 4 31B and Qwen 3.8-27B. Whisper, BGE-M3 and Kokoro are also available through SovAPI.

Direct model call

Gemma 4 31B
"model": "gemma-4-31b"
Text, vision and structured extraction.

Direct model call

Qwen 3.8-27B
"model": "qwen3.8-27b"
Code, reasoning, vision and tool calling.

Benchmarks

Faster

The inference race

At 256 output tokens, the Qwen comparison estimates 1.13 s on Sovinfra and 11.23 s on Prime Intellect (≈9.9×).

View the animated speed comparison
  1. Sovinfra · Qwen 3.8-27B1.13 s

    137 ms first token · 255.62 tokens/s

  2. Sovinfra Dedicated · Gemma 4 31B≤ 2.67 s

    120 ms first token · 100 tokens/s · Dedicated thresholds, excluding queue time

  3. Anthropic · Claude Sonnet 52.71 s

    723 ms first token · 128.47 tokens/s

  4. Anthropic · Claude Haiku 4.53.33 s

    400 ms first token · 87 tokens/s · OpenRouter

  5. Anthropic · Claude Opus 4.84.44 s

    850 ms first token · 71 tokens/s · OpenRouter

  6. Prime Intellect · DeepSeek V4 Flash 07315.35 s

    880 ms first token · 57 tokens/s · OpenRouter

  7. OpenAI · GPT-5.6 Luna5.65 s

    1654 ms first token · 63.78 tokens/s

  8. Prime Intellect · Gemma 4 31B10.14 s

    330 ms first token · 26 tokens/s · OpenRouter

  9. Prime Intellect · Qwen 3.8-27B11.23 s

    1420 ms first token · 26 tokens/s · OpenRouter

Estimates at equal output length, combining first-token latency and generation. Sovinfra Arena and OpenRouter observations were taken under different conditions. On playback, each row starts at its first token, then advances at its generation rate. Playback time is scaled for display.

Prime’s DeepSeek V4.1 Flash 1017 observation of 99 tokens/s needs a paired first-token measurement before a total can be calculated.

Cheaper

≈1,380× lower effective cost per billion tokens

$3 on the Sovinfra Shared API plan vs. $4,140.63 for Claude Opus 4.8 via Anthropic on OpenRouter, per billion total tokens.

At a 15:1 input-to-output ratio with 50% of input cached, the comparison basket contains 1B input and 66.67M output tokens. It costs $4,416.67 on OpenRouter. That effective rate assumes 4B total tokens are used on the $12/month plan. If the quota is weekly, this means using 1B in each of four weeks. Figures exclude tax.

View unit prices and the full cost comparison

Sovinfra.Cost

Sovinfra’s intelligence per dollar

Cost comparison

Sovinfra pricing

Same basket in both views · 1B input tokens · ~66.67M output tokens · 15:1 input-to-output ratio

Share of input tokens read from cache

50% CACHE · PRICING COMPARISON (USD)
Provider / model Rates in USD per million tokensPrice / cost
Sovinfra · Qwen / Gemma$12/month · $3 per billion tokens at full usageShared API · performance varies with load. Dedicated thresholds apply to the agreed configuration and workload.$3.00per 1B tokens
Prime · Gemma 4 31BInput $0.2 · Cache $0.1 · Output $0.335OpenRouter · 0.33–0.47 s latency · 19–26 tok/s$172.33
Prime · DeepSeek V4.1 Flash 1017Input $0.0437 · Cache $0.0024 · Output $0.7OpenRouter · 93 tok/s · latency unavailable in this reading$69.72
Prime · DeepSeek V4 FlashInput $0.44 · Cache $0.014 · Output $1.32OpenRouter · 0731 variant · 0.88 s latency · 57 tok/s$315.00
Prime · Qwen 3.8 27BInput $0.45 · Cache $0.05 · Output $3.2OpenRouter · 0.93–1.42 s latency · 11–26 tok/s$463.33
Prime · Claude Haiku 4.5Input $1 · Cache uses input rate · Output $5No Prime-specific OpenRouter measurement found.$1,333.33
Prime · Claude Sonnet 5Input $2 · Cache uses input rate · Output $10No Prime-specific OpenRouter measurement found.$2,666.67
Prime · Claude Opus 4.8Input $5 · Cache uses input rate · Output $25No Prime-specific OpenRouter measurement found.$6,666.67
Anthropic via OpenRouter · Claude Haiku 4.5Input $1 · Cache $0.1 · Output $5OpenRouter · 0.40 s latency · 87 tok/s$883.33
Anthropic via OpenRouter · Claude Sonnet 5Input $2 · Cache $0.2 · Output $10OpenRouter · 0.95 s latency · 76 tok/s$1,766.67
Anthropic via OpenRouter · Claude Opus 4.8Input $5 · Cache $0.5 · Output $25OpenRouter · 0.85 s latency · 71 tok/s$4,416.67

Basket estimates use the same token volumes and exclude taxes. The Sovinfra Shared Plan row shows its effective rate per billion total tokens at full use of the $12 monthly plan. Per-Token Pricing compares reference rates without subscription fees. Standard input rates apply where no cached rate is listed. Prime rates from 14 November 2026 screenshots. Anthropic rates are those listed by OpenRouter for the Anthropic endpoints. OpenRouter performance figures are observations for the linked endpoints and vary over time. OpenRouter observations and Sovinfra Dedicated guarantees use different conditions.

Workflow intelligence

Intelligence per dollar

We had to extend the x-axis.

At the same Qwen 3.8-27B reference score, the estimated cost per task is ≈$0.0065 on Shared Plan at full usage, ≈$0.21 at Sovinfra per-token rates and ≈$0.937 on Prime Intellect.

Estimated cost per benchmark task versus Intelligence Index. At the Qwen reference score of 34, Sovinfra Shared Plan is about $0.0065 at full usage, Sovinfra Per-Token is $0.21, Prime Intellect is $0.937, and the Artificial Analysis reference is $1.01.
Sovinfra and Prime use the Qwen reference score and the same estimated token volumes (50% of input cached, 15:1 input to output). Shared Plan uses an effective rate of $3 per billion total tokens at full use of the $12 monthly plan. Qwen and Luna benchmark references: Artificial Analysis v4.3.2. Other model positions are indicative. Qwen 3.8-27B model page.

Results published by the model developers, according to their evaluation protocols.

Qwen researchGoogle Gemma

Try Sovinfra

API Pricing.

SHARED API

1 BILLION TOKENS PER WEEK

$12 USD per month per account, before tax

$14.40 USD per month with 20% French VAT, when applicable. Your total depends on your billing location and is shown before payment.

For your agents, tools, and applications.

  • Professional use, integration into commercial applications, and resale of API keys are permitted.
  • All API keys on the account share the weekly allowance of one billion tokens. Cached input, uncached input, and output tokens all count toward the same allowance.
  • Usage beyond the allowance is billed per token at the rates below.
  • 1 user per account. Sharing login credentials is prohibited.
  • You can cancel at the end of any monthly billing period.

Performance

< 120 ms time to first token, excluding queue time

> 100 tokens/s generation speed per request

Throughput and latency vary with shared load. Individual performance is not guaranteed.

Per-token rates

Shared API · Input and output beyond the shared weekly allowance.

Rates per million tokens, excluding tax. All prices are in US dollars.
Uncached input$0.12
Cached input$0.04
Output$0.38

Included models and capabilities

  • Qwen3.8-27B · FP8 on H200.Code, reasoning, vision, and tool calling.
  • Gemma 4 31B · FP8 with BF16 KV cache.Text, vision, and data extraction.
  • BGE-M31,024-dimensional embeddings and semantic search.
  • Whisper Large v3Multilingual transcription of WAV, MP3, and OGG files.
  • KokoroText-to-speech in French, US English, and UK English.
SubscribeGet your free API key →

Start with 1 billion free tokens. No credit card required.

14-day trial · 100 million tokens per day · Up to 2 concurrent requests processed per account.
Up to 160,000 input tokens per Qwen or Gemma request.

DEDICATED API

Reserved capacity

Per-token billing

Standard guarantees

  • < 120 msTime to first token, excluding queue time
  • > 100 tokens/sGeneration speed per request

For production applications that require reserved GPU capacity.

  • Reserved GPU capacity, with sizing tailored to your workload on request.
  • Input and output are billed from the first token. No subscription is required.
  • Additional throughput and latency guarantees are available on request, depending on your workload.
  • An uptime SLA of 99.95% or higher is available on request.

Standard guarantees apply to the agreed configuration and workload.

Tell us about your workload →
Full plan details

Qwen3.8-27B · FP8 on H200.

Gemma 4 31B · FP8 with BF16 KV cache on H200.

Qwen and Gemma accept text and image inputs. Context window of 262,144 tokens.

No per-account limit on concurrent or queued requests.

Caching speeds up repeated requests and reduces costs beyond the allowance.

Your data is not used to train models. Requests and responses are not retained in logs.

Requests are processed on GPUs located in Europe with a named hosting provider.

Model versions and quantization are disclosed.

Usage dashboard and email support.

EU hosting · Your data is excluded from training · OpenAI-compatible API

Real-world applications

Built for Real-World Work.

European AI inference for software publishers, business automation and industrial applications.

Business automation

AI agents, CRM workflows, customer support and business process automation.

Recommended model
Qwen 3.8-27B
Benchmark reference
AutomationBench-AA

Document intelligence

Extract, analyse and structure information from contracts, invoices, reports and business documents.

Recommended models
Gemma 4 31B and Qwen 3.8-27B
Benchmark reference
GDP.pdf (Qwen 3.8-27B)

Industrial operations

Production analytics, operational reporting, data processing and decision-support workflows.

Recommended models
Qwen 3.8-27B and Gemma 4 31B

Flexible model settings

Your model. Your settings.

Start with production defaults. Adjust reasoning depth and image processing for more demanding workloads.

Qwen 3.8-27B

Reasoning on demand

Three reasoning levels are available: none, medium and xhigh. The default is none (reasoning disabled). For more demanding tasks, increase the reasoning level in your API request.

"reasoning_effort": "medium"
"reasoning_effort": "xhigh"

Higher reasoning levels may increase token consumption and latency.

Gemma 4 31B

Adjust image processing

The default image-processing budget is 560. For complex invoices, receipts and visual documents, increase it to 1120 in your API request.

"mm_processor_kwargs": { "max_soft_tokens": 1120 }

Higher image-processing budgets may increase token consumption and processing time.

Built for European workloads

Security. Compliance.

01

EU-hosted

All inference runs on European GPUs. No data leaves the EU.

02

GDPR native

No training on user data. End-to-end encryption. DPA available.

03

No retention

Requests are processed in the EU and are not retained or used to train models.

04

Auditable

Our infrastructure stack is open for inspection on Codeberg. Full transparency on security and routing.

Sovinfra | Inference built around your workload.