Loading the catalogue…

GPT-4-turbo API price: €9.07/1M input tokens · Stav AI

Inference Providers

Inference Provider	Sovereignty	Runs in	In / out per 1M	Cache r / w	Context · max out	Speed	Uptime 7 d	Status	Pin
OpenAI	L0		€9.07 / 27.22	—	128K · 4.1K	1.3 s · 28 tok/s	No data	Online

Facts from provider documents, last confirmed 26 Sept 2026.Prices checked daily.* Prices exclude VAT.

What it's good for

21 use cases classified by the curator. The note says why.

Strong

Customer Support Chatbotstrong

High safety benchmark (0.997) and tool-calling support suit customer-facing chatbot deployments requiring guardrail reliability.

Document Q&Astrong

128K context window supports full-document ingestion for accurate Q&A over long source material.

Image Q&Astrong

Native vision capability enables direct question-answering over image inputs.

Policy Q&Astrong

0.997 on the Anthropic red-team safety set plus 128K context makes this a strong fit for compliance-sensitive policy Q&A workloads.

Research Synthesisstrong

Large 128K context window enables synthesis across multiple long documents in a single pass.

Summarisationstrong

128K context handles long transcripts/reports for summarisation without chunking.

Acceptable — read the note

Chart Extractionacceptable

Vision capability allows chart interpretation, but no dedicated chart-extraction benchmark is documented.

Document OCRacceptable

Vision input supports document image parsing, though not purpose-built or benchmarked as an OCR specialist.

Email Draftingacceptable

Solid general instruction-following makes it a competent choice for routine business drafting tasks.

Function-Calling Appsacceptable

Native tool-calling capability suits structured function-calling application patterns.

Long-Form Writingacceptable

128K context supports long-form drafting, though the family is positioned as a general instruct model rather than a creative specialist.

Screenshot Understandingacceptable

Text+image input supports screenshot interpretation tasks at interactive latency.

Weak or not recommended

Agentic Codingnot recommended

Weak coding benchmarks (21.5 coding index, 37.3 LiveCodeBench) rule this out for autonomous coding-agent workflows despite tool-calling support.

Bug Diagnosisweak

Low LiveCodeBench score (37.3) indicates weaker code-reasoning capability relevant to bug diagnosis.

Code Reviewweak

Family posts only a 21.5 coding index and 37.3 LiveCodeBench score, materially behind code-tuned model releases.

Speech Synthesisnot recommended

No speech synthesis capability documented for this text+image instruct model.

Test Generationweak

Coding-heavy workloads lag newer code-tuned releases per the family's 21.5 coding index.

Transcriptionnot recommended

No documented audio input/output capability — model card lists only tool-calling and vision.

Get started

curl https://api.stav.ai/v1/chat/completions \
  -H "Authorization: Bearer $STAV_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4-turbo",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

→ X-Stav-Request-Id · X-Stav-Model · X-Stav-Provider → X-Stav-Sovereignty-Level — the rung the request was served at

gpt-4-turbo

Use this model

Capabilities

Inference Providers

What it's good for

How good is it

Strengths against the fieldi

Parameters this model honours

Get started