This page lists all models supported by the Archia server. Use the Model ID value when specifying a model in the
Responses API or in agent configuration files.
Remote models are hosted by their respective providers and require API keys to be configured.
Model ID Display Name Description Capabilities
gpt-5.2GPT-5.2 OpenAI’s most advanced model nlp, vision, tools, reasoning gpt-5.1GPT-5.1 Improved GPT-5 with better reasoning, coding, and vision nlp, vision, tools, reasoning gpt-5GPT-5 Flagship general model for coding, reasoning, and agentic tasks nlp, vision, tools, reasoning gpt-5-miniGPT-5 mini Fast, affordable small model for focused tasks nlp, vision gpt-5-nanoGPT-5 nano Smallest GPT-5 variant; optimized for cost and latency nlp gpt-4.1GPT-4.1 Improved GPT-4 with better reasoning, coding, and vision nlp, vision, tools, reasoning gpt-4oGPT-4o Omni multimodal model; strong general performance and tool calling nlp, vision, tools gpt-4o-miniGPT-4o mini Fast, affordable small model for focused tasks nlp, vision o4-miniOpenAI o4-mini Latest small o-series model optimized for efficient reasoning and coding nlp, vision, reasoning
Model ID Display Name Description Capabilities
gemini-3-pro-previewGemini 3.0 Pro Google’s flagship Gemini 3.0 Pro model with advanced reasoning and multimodal capabilities nlp, vision, tools, reasoning gemini-2.5-proGemini 2.5 Pro Google’s high-performance model with strong reasoning and multimodal capabilities nlp, vision, tools, reasoning gemini-2.5-flashGemini 2.5 Flash Google’s lightweight model optimized for speed and efficiency nlp, vision, tools, reasoning
Model ID Display Name Description Capabilities
claude-opus-4-5-20251101Opus 4.5 Claude Opus 4.5 Premium model combining maximum intelligence with practical performance nlp, vision, tools, reasoning claude-opus-4-1-20250805Opus 4.1 Latest flagship Claude with strongest reasoning/coding and vision nlp, vision, reasoning, tools claude-sonnet-4-5-20250929Sonnet 4.5 Claude 4.5: improved reasoning, coding, and multimodal capabilities with lower latency nlp, vision, reasoning, tools claude-sonnet-4-20250514Sonnet 4 High-performance Claude with 200k–1M context and vision nlp, vision, reasoning, tools claude-3-7-sonnet-20250219Sonnet 3.7 Fast, capable Claude with extended thinking options nlp, vision, reasoning, tools claude-haiku-4-5-20251001Haiku 4.5 Anthropic’s fastest model with near-frontier intelligence nlp, vision, tools claude-3-5-haiku-20241022Haiku 3.5 Small/fast Claude for chat + basic vision nlp, vision, tools
Private models route through Archia’s secure infrastructure with zero data retention guarantees. These are ideal for
sensitive workloads where data privacy is critical.
Model ID Display Name Description Capabilities
priv-claude-opus-4-5-20251101Opus 4.5 (Private) Claude Opus 4.5 with no data retention nlp, vision, tools, reasoning priv-claude-opus-4-1-20250805Opus 4.1 (Private) Latest Claude Opus with no data retention nlp, vision, reasoning, tools priv-claude-sonnet-4-5-20250929Sonnet 4.5 (Private) Claude 4.5 with no data retention nlp, vision, tools, reasoning priv-claude-sonnet-4-20250514Sonnet 4 (Private) Claude Sonnet 4 with no data retention nlp, vision, reasoning, tools priv-claude-3-7-sonnet-20250219Sonnet 3.7 (Private) Fast, capable Claude with no data retention nlp, vision, reasoning, tools priv-claude-haiku-4-5-20251001Haiku 4.5 (Private) Anthropic’s fastest model with no data retention nlp, vision, tools, reasoning priv-claude-3-5-haiku-20241022Haiku 3.5 (Private) Fast/cheap Claude with no data retention nlp, vision, tools
Model ID Display Name Description Capabilities
archia::openai/gpt-oss-20bGPT OSS 20B Open-source 20B parameter model hosted on Archia infrastructure vision, nlp, reasoning archia::openai/gpt-oss-120bGPT OSS 120B Open-source 120B parameter model with enhanced capabilities vision, nlp, reasoning archia::moonshotai/kimi-k2-instruct-0905Kimi K2 Open source model from Moonshot AI (~1 trillion parameters) vision, nlp, reasoning
Local models run directly on your machine using llama.cpp. They require downloading the model weights and sufficient
system memory.
Model ID Display Name Description Capabilities Size Min Memory
gpt-oss-20b-F16.ggufgpt-oss-20b Open-source 20B parameter model optimized for performance vision, nlp, reasoning 12 GB 32 GB gpt-oss-120b-F16.ggufgpt-oss-120b Open-source 120B parameter model for complex reasoning tasks vision, nlp, reasoning 60 GB 100 GB Qwen3-4B-Instruct-2507-F16.ggufqwen3-4b Open-source 4B instruction-tuned model for general tasks nlp 8 GB 32 GB Qwen3-4B-Q5_K_S.ggufqwen3-4b-think 4B model with enhanced thinking, 5-bit quantized nlp 8 GB 16 GB Qwen3-32B-Q5_K_S.ggufqwen3-32b-think 32B reasoning model with advanced thinking depth nlp 22 GB 32 GB
Models are tagged with capabilities that indicate their supported features:
Capability Description
nlpNatural language processing (text generation, understanding) visionImage understanding and analysis toolsFunction/tool calling support reasoningExtended thinking/reasoning support (see Reasoning )
{
"model": "claude-sonnet-4-5-20250929",
"input": "Hello, how can you help me today?"
}
In your agent TOML configuration file:
[assistant]
model = "gpt-5.2"
system_prompt = "You are a helpful assistant."
For sensitive workloads, use the private model variants:
{
"model": "priv-claude-sonnet-4-5-20250929",
"input": "Process this confidential document..."
}
Use Case Recommended Models
General tasks gpt-5.2, claude-sonnet-4-5-20250929, gemini-2.5-proFast responses gpt-5-mini, claude-haiku-4-5-20251001, gemini-2.5-flashComplex reasoning claude-opus-4-5-20251101, gpt-5.2 with reasoning, o4-miniCost optimization gpt-5-nano, claude-3-5-haiku-20241022, gpt-4o-miniPrivacy-sensitive priv-claude-* models, archia::* modelsOffline/local gpt-oss-20b-F16.gguf, Qwen3-* modelsVision tasks Any model with vision capability