AI Models

Access cutting-edge AI models through our unified API. From language understanding to complex reasoning.

Available Language Models

Live Now

Alibaba

Qwen3-32B
Alibaba
33k tokensMax Output 4k tokens

Qwen3 32B — multilingual model with strong Chinese/English and reasoning

Aussie SovereignReasoningTools

DeepSeek

DeepSeek-R1-0528
DeepSeek
131k tokensMax Output 7k tokens

First-generation RL‑trained reasoning model with chain‑of‑thought (o1‑class)

Aussie SovereignReasoning
DeepSeek-R1-Distill-Llama-70B
DeepSeek
131k tokensMax Output 4k tokens

Distilled Llama 70B variant of DeepSeek R1 with strong reasoning

Aussie SovereignReasoning
DeepSeek-V3.1
DeepSeek
131k tokensMax Output 7k tokens

Hybrid MoE model with distilled R1 reasoning; excels at math, code, agents

Aussie SovereignReasoningTools
DeepSeek-V3.1-Terminus
DeepSeek
131k tokensMax Output 7k tokens

DeepSeek V3.1 Terminus — enhanced reasoning with reliable function calling

Aussie SovereignReasoningTools
DeepSeek-V3-0324
DeepSeek
131k tokensMax Output 7k tokens

DeepSeek V3 (MoE) focused on fast direct outputs; 128k context, tools

Aussie SovereignTools

HUMAIN

ALLaM-7B-Instruct-preview
HUMAIN
4k tokensMax Output 32k tokens

ALLaM 7B Instruct — compact instruction-tuned model for general tasks

Aussie Sovereign

Meta

Llama-3.3-Swallow-70B-Instruct-v0.4
Meta
131k tokensMax Output 3k tokens

Meta Llama 3.3 70B Instruct — multilingual with 128k context (GQA‑optimized)

Aussie SovereignTools
Llama-4-Maverick-17B-128E-Instruct
Meta
131k tokensMax Output 4k tokens

Meta Llama 4 Maverick 17B — native multimodal MoE (128 experts), 128k context

Aussie SovereignToolsVision
Meta-Llama-3.1-8B-Instruct
Meta
16k tokensMax Output 4k tokens

Meta Llama 3.1 8B — efficient multilingual model with 128k context and GQA

Aussie SovereignTools

OpenAI

gpt-oss-120b
OpenAI
131k tokensMax Output 131k tokens

Open‑weight MoE reasoning model (117B/5.1B active), 131k context, CoT & tools

Aussie SovereignReasoningTools

scx.ai

magpie-small
scx.ai
131k tokensMax Output 131k tokens

Experimental Australian sovereign fine‑tuned model optimized for Australian English

Aussie SovereignReasoning

Available Embedding Models

Live Now

Mistral

E5-Mistral-7B-Instruct
Mistral
33k tokens4.1k dims

Compact 7B embedding model for semantic search and retrieval

Aussie SovereignEmbedding

Available Audio Models

Live Now

OpenAI

Whisper-Large-v3
OpenAI
4k tokensMax Output 4k tokens

Multilingual ASR for transcription and translation (99+ languages)

Aussie SovereignTranscription

On-Demand Models

Available for Enterprise

We also provide access to a wider range of specialized models for enterprise clients with specific needs. Contact us for more information on enabling these models for your organization.

Model FamilyContextDescriptionCapabilities
Google Gemma Series
Up to 8k tokensHigh-performance model based on Google AI technology for ultra-fast inference.
Ultra-fast InferenceLightweight Deployment
Voice & Audio Models
Real-time processingWhisper and other leading voice models for transcription and processing.
Speech-to-TextAudio AnalysisVoice Synthesis
Embedding Models
Vector generationAdvanced embedding models for semantic search and RAG applications.
Document SearchSimilarity MatchingKnowledge Retrieval
Video & Multimodal
Multi-format inputNext-generation models for video analysis and multimodal understanding.
Video ProcessingImage AnalysisCross-modal Reasoning
Australian Domestic Models
CustomText, voice, and video models created by Australia's leading AI experts.
TextVoiceVideoSovereign AI