Powered by

# Model Serving - Active Models

This is the list of AI models currently live and available on the Phoeniqs Model Service. All models are served on Phoeniqs infrastructure through an OpenAI-compatible API and are ready for production use.

For real-time health and availability, see the model status dashboard.

# Active Models

Input and Output (chf): credits per million tokens. Context window: maximum tokens per API call.

# Chat

Instruction-following models for assistants, dialogue, and general text tasks.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-deepseek-v32 deepseek-ai/DeepSeek-V3.2 MIT 0.6156 1.8469 163'840 Optimized for Reasoning chat completions.
inference-gpt-oss-120b openai/gpt-oss-120b Apache 2.0 0.1154 0.4617 131'072 Optimized for powerful reasoning, agentic tasks, and versatile developer use cases.
inference-granite-33-8b ibm-granite/granite-3.3-8b-instruct Apache 2.0 0.1539 0.1539 32'768 Optimized for Reasoning and instruction-following capabilities.
inference-mistral-v03-7b mistralai/Mistral-7B-Instruct-v0.3 Apache 2.0 0.1539 0.1539 32'768 Optimized for multilingual dialogue use cases.

# Reasoning

Models for deliberate, step-by-step thinking and chain-of-thought tasks.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-deepseek-v4-flash deepseek-ai/DeepSeek-V4-Flash-0731 MIT 0.1462 0.3925 1'000'000 Optimized for thinking and reasoning with extended context.
inference-glm5 zai-org/GLM-5.2 MIT 1.077 3.386 131'072 Optimized for agentic engineering and long-horizon coding. Designed for sustained performance across extended agentic sessions with stronger coding, terminal, and software-engineering task capabilities.
inference-qwen3-8b RedHatAI/Qwen3-8B-quantized.w4a16 Apache 2.0 0.0269 0.1062 40'960 Optimized for thinking and reasoning.
inference-qwq-32b RedHatAI/QwQ-32B-quantized.w8a8 Apache 2.0 0.9234 0.9234 32'768 Optimized for thinking and reasoning.

# Multimodal

Vision-language models for image understanding, analysis, and text generation from visual input.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-apertus-v15-70b swiss-ai/Apertus-v1.5-70B Apache 2.0 0.6195 2.2201 262'144 Optimized for multilingual dialogue, tool use, and agent workflows.
inference-gemma4-31b RedHatAI/gemma-4-31B-it-FP8-block Apache 2.0 0.118 0.325 131'072 Optimized for handling text and image input and generating text output.
inference-granite-vision-2b ibm-granite/granite-vision-3.2-2b Apache 2.0 0.0770 0.0770 8'192 Optimized for compact and efficient vision-language model
inference-llama4-maverick RedHatAI/Llama-4-Maverick-17B-128E-Instruct-quantized.w4a16 Llama 4 Community License (custom) 0.2693 1.0773 1'048'576 Optimized for text and multimodal experiences. Max images per prompt is 4 and no video prompts
inference-llama4-scout-17b RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16 Llama 4 Community License 0.1924 0.6387 62'256 Optimized for text and multimodal experiences.
inference-qwen3-vl-235b RedHatAI/Qwen3-VL-235B-A22B-Instruct-FP8-dynamic Apache 2.0 0.7003 2.0000 128'000 Optimized for text and multimodal experiences.

# Embedding

Convert text into vector representations for semantic search and RAG retrieval.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-bge-m3 BAAI/bge-m3 MIT 0.4309 --- — Optimized for Embeddings and sparse retrieval with support for Multi-Functionality, Multi-Linguality, and Multi-Granularity.
inference-granite-emb-278m ibm-granite/granite-embedding-278m-multilingual Apache 2.0 0.0770 --- — Optimized for Embeddings.
inference-qwen3-embedding-8b Qwen/Qwen3-Embedding-8B Apache 2.0 0.0540 --- — Optimized for text embeddings and retrieval across 100+ languages.

# Reranker

Score and reorder retrieved passages by relevance to improve search quality.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-bge-reranker BAAI/bge-reranker-v2-m3 Apache 2.0 0.0077 --- — Optimized for Reranker to get relevance score.
inference-qwen3-reranker-8b Qwen/Qwen3-Reranker-8B Apache 2.0 0.0080 --- — Optimized for reranking to get relevance scores across 100+ languages.

# OCR

Extract text from images and scanned documents using optical character recognition.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-deepseek-ocr inference-deepseek-ocr MIT 0.3848 1.5391 — Optimized for Contexts Optical Compression.

# Vision-language

Parse complex document layouts and extract structured content from PDFs and files.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-miner-u25 opendatalab/MinerU2.5-2509-1.2B AGPL 3.0 0.38 0.23 8'192 Optimized for document parsing that achieves state-of-the-art accuracy with high computational efficiency.

# Speech to Text

Transcribe spoken audio into text, including long-form and multilingual speech.

Model Name Model License type Input (chf) Output (chf) Context window Description
inference-whisper-large-v3 openai/whisper-large-v3 Apache 2.0 0.006 per minute NA — Optimized for automatic speech recognition (ASR) and speech translation. WhisperX model support high throughput, ASR, long audio files support, speaker diarization & attribution and word-level alignments.
inference-flix-swissgerman-full Flix-AI/flix-swissgerman-full Apache 2.0 0.006 per minute NA — Optimized for Swiss German automatic speech recognition (ASR). Transcribes Swiss German dialect speech into Standard German text via WhisperX.

Model Updates and Deprecation Disclaimer We reserve the right to modify, upgrade, or replace any AI models used in our services at any time. This may include deprecating older models and introducing newer versions as we deem necessary to maintain performance, security, and service quality. While we aim to provide notice when feasible, changes may occur without prior notification.

NOTE Pricing is subject to change at our discretion.

# Active Models by Use Case

The groups below mirror the Active Models sections: every live model appears exactly once, grouped by primary workload.

  • Chat, assistants, and general text (instruction-following)
    Multilingual dialogue, coding help, and broad assistant-style tasks: DeepSeek V3.2, GLM 4.6 (chat; deployed on demand), GPT OSS 120B, Granite 3.3 8B, Mistral 7B Instruct v0.3.

  • Reasoning and chain-of-thought
    Deliberate, step-heavy tasks: DeepSeek V4 Flash, GLM 5, Qwen3 8B, QwQ 32B.

  • Multimodal (images and text)
    Image+text input and text output, or compact vision-language: Apertus v1.5 70B, Gemma 4 31B, Granite Vision 3.2 2B, Llama 4 Maverick, Llama 4 Scout 17B, Qwen3 VL 235B.

  • RAG (retrieval-augmented generation)
    Embeddings, reranking, and retrieval stacks: BGE M3 (embedding), BGE Reranker v2 M3, Granite Embedding 278M, Qwen3 Embedding 8B, Qwen3 Reranker 8B.

  • OCR, layout, and document parsing
    Optical compression, layout, and structured document workflows: DeepSeek OCR, MinerU 2.5.

  • Speech
    Transcription and speech-centric workflows: Whisper Large v3, Flix Swiss German Full.

  • File conversion To transform files from one format to another, use the Docling service hosted by PHOENIQS Model Service. You can access it via the REST API endpoint: https://maas.phoeniqs.com/v1/convert/file. For implementation details, see the sample API calls.

Agents, tool calling, and orchestration usually pick one primary LLM from the chat or reasoning groups, then add other pieces only when the workflow needs them: multimodal models for image inputs; OCR or MinerU to extract text or structure from documents; and RAG (embedding, reranking) when the agent must retrieve relevant passages from content you have already chunked and indexed.

# Active Models by Risk

Risk Level Typical Models Main Risks
Low Risk Embedding models, OCR, translation, sentiment analysis Limited misuse potential
Medium Risk General LLMs, code generation, image generation Hallucinations, insecure code, copyright issues
High Risk Unmoderated LLMs, NSFW models, unsafe diffusion models, executable pickle-based models Malware, harmful content, prompt injection, data leakage

# Using the models

Looking for ready-to-run examples? See the Model Service Guides:

maas ai model serving

© Copyright 2026. All rights reserved.