Model Catalog

Every model below is reachable through the same OpenAI-compatible endpoint. For each model, you see in which jurisdictions it can run and which provider serves it — your policy decides which lane is taken per request.

EU Sovereign
EU Hosted
Global

Frontier chat

Flagship multimodal models for the hardest workloads.

12 models
Mistral Large 3128K
chat · vision · tools · reasoning
  • 🇫🇷 France·Mistral AI
Llama 3.1 405B Instruct131K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
  • 🇩🇪 Germany·IONOS
Qwen 3.5 397B256K
chat · code · vision
  • 🇫🇷 France·Scaleway
Qwen 3 235B Instruct262K
chat
  • 🇫🇷 France·Scaleway
Qwen 3 VL 235B218K
chat · vision · tools
  • 🇩🇪 Germany·STACKIT
Hermes 4 405B128K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
Nemotron 3 Super 120B1M
chat · tools
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
GPT-5.4 Pro128K
chat · vision · tools · reasoning
  • 🇸🇪 Sweden Central·Azure OpenAI
GPT-5128K
chat · vision · tools
  • 🇸🇪 Sweden Central·Azure OpenAI
Claude Opus 4.7200K
chat · vision · tools · reasoning
  • 🇸🇪 Stockholm + 🇪🇺 EU·AWS Bedrock
Claude Opus 4.5200K
chat · vision · tools · reasoning
  • 🇸🇪 Stockholm + 🇪🇺 EU·AWS Bedrock
Gemini 3.1 Pro1M
chat · vision · tools · reasoning
  • 🇺🇸 United States·Google AI

Workhorse chat

Production-grade general-purpose models at mid-size.

21 models
Llama 3.3 70B Instruct128K
chat · tools
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius(fast lane)
  • 🇫🇷 France·Scaleway
  • 🇩🇪 Germany·IONOS
  • 🇩🇪 Germany·STACKIT
Qwen 2.5 72B Instruct131K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
Qwen 3 32B131K
chat · tools
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
gpt-oss-120b128K
chat · reasoning · tools
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
  • 🇫🇷 France·Scaleway
  • 🇩🇪 Germany·IONOS
  • 🇩🇪 Germany·STACKIT(131K context)
Hermes 4 70B128K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
GLM 5128K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
MiniMax M2.5200K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
DeepSeek V3 032464K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
Gemma 3 27B IT128K
chat · vision
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
  • 🇫🇷 France·Scaleway
  • 🇩🇪 Germany·STACKIT(37K context)
Mistral Medium 3.1128K
chat · vision · tools
  • 🇫🇷 France·Mistral AI
Mistral Small 3.1128K
chat · vision · tools
  • 🇫🇷 France·Mistral AI
Mistral Small 3.2 24B128K
chat · vision
  • 🇫🇷 France·Scaleway
Mistral Small 24B128K
chat
  • 🇩🇪 Germany·IONOS
GPT-5 Mini128K
chat
  • 🇸🇪 Sweden Central·Azure OpenAI
GPT-4.1128K
chat · vision · tools
  • 🇸🇪 Sweden Central·Azure OpenAI
GPT-4.1 Mini128K
chat
  • 🇸🇪 Sweden Central·Azure OpenAI
GPT-4o128K
chat · vision · tools
  • 🇸🇪 Sweden Central·Azure OpenAI
Claude Sonnet 4.5200K
chat · vision · tools
  • 🇸🇪 Stockholm + 🇪🇺 EU·AWS Bedrock
Claude 3.7 Sonnet200K
chat · vision · tools
  • 🇸🇪 Stockholm + 🇪🇺 EU·AWS Bedrock
Gemini 3.1 Flash1M
chat · vision · tools · fast
  • 🇺🇸 United States·Google AI
Gemini 3 Flash1M
chat · vision · tools · preview
  • 🇺🇸 United States·Google AI

Reasoning

Thinking-style models for multi-step problem solving.

7 models
Magistral Medium40K
reasoning
  • 🇫🇷 France·Mistral AI
DeepSeek R1 052864K
reasoning
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius(fast lane)
DeepSeek R1 Distill Llama 70B128K
chat · reasoning
  • 🇫🇷 France·Scaleway
QwQ 32B131K
reasoning
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
Qwen 3 Next 80B Thinking131K
reasoning
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
o3200K
reasoning
  • 🇸🇪 Sweden Central·Azure OpenAI
o3 Mini200K
reasoning
  • 🇸🇪 Sweden Central·Azure OpenAI

Code & developer

Code completion, FIM, and agentic coding.

5 models
Codestral 25.08256K
code · FIM
  • 🇫🇷 France·Mistral AI
Devstral Medium128K
code · agents
  • 🇫🇷 France·Mistral AI
Devstral 2 123B128K
chat · code
  • 🇫🇷 France·Scaleway
Qwen 3 Coder 30B256K
chat · code
  • 🇫🇷 France·Scaleway
Qwen 3 Coder Next 80B
code
  • 🇩🇪 Germany·IONOS

Vision & multimodal

Image-in models where vision is the primary capability.

4 models
Pixtral Large128K
vision · chat
  • 🇫🇷 France·Mistral AI
Pixtral 12B128K
vision · chat
  • 🇫🇷 France·Scaleway
Qwen 2.5 VL 72B Instruct131K
vision · chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
Holo2 30B128K
vision · chat
  • 🇫🇷 France·Scaleway

Compact & edge

Small open-weight models for high-throughput / low-latency tasks.

6 models
Ministral 8B128K
chat · tools
  • 🇫🇷 France·Mistral AI
Open Mistral Nemo128K
chat · open-weight
  • 🇫🇷 France·Mistral AI
Mistral Nemo Instruct128K
chat
  • 🇫🇷 France·Scaleway
  • 🇩🇪 Germany·IONOS
Llama 3.1 8B Instruct131K
chat
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius
  • 🇫🇷 France·Scaleway
  • 🇩🇪 Germany·IONOS
gpt-oss-20b131K
chat · reasoning · tools
  • 🇩🇪 Germany·STACKIT
Gemini 3.1 Flash-Lite1M
chat · fast
  • 🇺🇸 United States·Google AI

Embeddings

Vector embeddings for retrieval, classification and search.

9 models
mistral-embed8K
embeddings · 1024-dim
  • 🇫🇷 France·Mistral AI
codestral-embed8K
code embeddings
  • 🇫🇷 France·Mistral AI
Qwen 3 Embedding 8B32K
embeddings
  • 🇳🇱 NL · 🇫🇮 + 🇫🇷 inference·Nebius(8K context)
  • 🇫🇷 France·Scaleway
Qwen 3 VL Embedding 8B32K
embeddings
  • 🇩🇪 Germany·STACKIT
E5 Mistral 7B4K
embeddings · 1024-dim
  • 🇩🇪 Germany·STACKIT
BGE Multilingual Gemma 28K
embeddings · multilingual
  • 🇫🇷 France·Scaleway
bge-m3
embeddings · multilingual
  • 🇩🇪 Germany·IONOS
bge-large-en-v1.5
embeddings
  • 🇩🇪 Germany·IONOS
paraphrase-multilingual-mpnet-base-v2
embeddings · multilingual
  • 🇩🇪 Germany·IONOS

Document & OCR

Layout-aware document parsing and optical character recognition.

2 models
Mistral OCR
OCR
  • 🇫🇷 France·Mistral AI
LightOnOCR 2
OCR
  • 🇩🇪 Germany·IONOS

Speech

Speech-to-text and text-to-speech.

3 models
Voxtral Small
ASR
  • 🇫🇷 France·Mistral AI
  • 🇫🇷 France·Scaleway(24B · 32K context)
Voxtral TTS
TTS
  • 🇫🇷 France·Mistral AI
Whisper Large v3
STT
  • 🇫🇷 France·Scaleway

Image generation

Text-to-image inference.

1 model
FLUX.1 [schnell]
image generation
  • 🇩🇪 Germany·IONOS

Per-token pricing, batch discounts, dedicated-endpoint options and DPA status per provider available on request. Catalogue refreshed monthly against each provider's public price list.

Request full price book

Provider Network

We aggregate LLM providers across the full spectrum behind a single endpoint. Every tier is accessible via the same API — TechVera routes to the right one based on your policy.

EU Sovereign

Providers operating exclusively under EU jurisdiction — no US parent company, no Cloud Act exposure.

Mistral AI
🇫🇷 France
Nebius Token Factory
🇳🇱 NL · 🇫🇮 + 🇫🇷 inference
Scaleway
🇫🇷 France
IONOS AI Model Hub
🇩🇪 Germany
STACKIT
🇩🇪 Germany — Schwarz Digits
EU Hosted

US hyperscalers in EU regions — opt-in only, with explicit CLOUD-Act labelling on every response.

Azure OpenAI
🇸🇪 Sweden Central Data Zone
AWS Bedrock
🇸🇪 Stockholm + 🇪🇺 EU regions
Global

Direct US / international providers. Available with explicit customer opt-in for non-sensitive workloads.

Google AI
🇺🇸 United States — Gemini 3.x

You operate LLM infrastructure in the EU?

We're actively onboarding sovereign and EU-hosted inference providers.

Become a provider partner