AI Deployment Profile Matrix
September 17, 2026 ยท View on GitHub
This file is generated from ir/axcore/data/provider-profiles.json. Do not edit it by hand. Capability values are deployment defaults; exact profile model rules and caller modelInfo overrides can refine them.
| Profile | Transport | Endpoint | Auth | Default capabilities | Model caveat | Official sources | Reviewed |
|---|---|---|---|---|---|---|---|
openai | openai-chat | https://api.openai.com/v1 | bearer | tools, stream, structured, modes native -> function -> json_object, thinking, images, audio, files, tiers standard/flex/priority | prefix gpt-6-astra: structured modes native -> json_object | source | 2026-08-17 |
openai-compatible | openai-chat | Required apiURL | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
openai-responses | openai-responses | https://api.openai.com/v1 | bearer | tools, stream, structured, modes native -> function -> json_object, thinking, images, audio, files, tiers standard/flex/priority | prefix gpt-6-astra | source | 2026-08-17 |
anthropic | anthropic-messages | https://api.anthropic.com | x-api-key | tools, stream, modes function, thinking, images | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
google-gemini | gemini-generate-content | https://generativelanguage.googleapis.com/v1beta | api-key-header | tools, stream, structured, modes native -> function, thinking, images, audio, files, tiers standard/flex/priority | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-17 |
webllm | webllm | Host runtime | none (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
azure-openai | openai-chat | resourceName + deploymentName | api-key-header | tools, stream, structured, modes native -> function, thinking, images, tiers standard/priority | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-17 |
deepseek | openai-chat | https://api.deepseek.com | bearer | tools, stream, modes function -> json_object | exact deepseek-v4-flash, deepseek-v4-pro: thinking, no native structured output, structured modes function, default thinking max, no forced tool choice; exact deepseek-reasoner: thinking, no native structured output, structured modes function, no forced tool choice | source | 2026-08-18 |
deepseek-responses | openai-responses | https://api.deepseek.com | bearer | tools, stream, modes function, thinking | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
meta | openai-responses | https://api.meta.ai/v1 | bearer | tools, stream, structured, modes native -> function -> json_object, thinking, images, audio, files | exact muse-image-1.0: no native structured output, structured modes ; exact muse-voice-transcribe-1.0: no native structured output, structured modes | source | 2026-09-03 |
meta-chat | openai-chat | https://api.meta.ai/v1 | bearer | tools, stream, structured, modes native -> function -> json_object, thinking, images, audio, files | Conservative defaults; use modelInfo for model-specific features | source | 2026-09-03 |
meta-messages | anthropic-messages | https://api.meta.ai/v1 | bearer | tools, stream, structured, modes native -> function, thinking, images, audio, files | Conservative defaults; use modelInfo for model-specific features | source | 2026-09-03 |
mistral | openai-chat | https://api.mistral.ai/v1 | bearer | tools, stream, structured, modes native -> function, images, audio, tiers standard/priority | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-17 |
cohere | openai-chat | https://api.cohere.ai/compatibility/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
grok | openai-chat | https://api.x.ai/v1 | bearer | tools, stream, structured, modes native -> function, images, audio, web search, tiers standard/priority | exact grok-4.6: thinking, structured modes native -> function, default thinking max; exact grok-4.5, grok-4.5-latest, grok-build-latest: thinking, structured modes native -> function, default thinking max; exact grok-4.3, grok-4.3-latest, grok-latest: thinking, structured modes native -> function, default thinking max; exact grok-3-mini, grok-3-mini-latest, grok-3-mini-beta, grok-3-mini-fast, grok-3-mini-fast-latest, grok-3-mini-fast-beta: thinking, default thinking low | source 1, source 2, source 3, source 4 | 2026-08-30 |
reka | openai-chat | https://api.reka.ai/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
together | openai-chat | https://api.together.xyz/v1 | bearer | tools, stream, structured, modes native -> function -> json_object | prefix deepseek-ai/DeepSeek-V4: thinking, no native structured output, structured modes function, default thinking max, no forced tool choice | source | 2026-08-18 |
openrouter | openai-chat | https://openrouter.ai/api/v1 | bearer | tools, stream, modes function, tiers standard/flex/priority | prefix deepseek/: thinking, no native structured output, structured modes function, default thinking max, no forced tool choice | source 1, source 2 | 2026-08-18 |
orcarouter | openai-chat | https://api.orcarouter.ai/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-19 |
fireworks | openai-chat | https://api.fireworks.ai/inference/v1 | bearer | tools, stream, structured, modes native -> function, tiers standard/priority | contains deepseek-v4: thinking, no native structured output, structured modes function, default thinking max, no forced tool choice | source 1, source 2 | 2026-08-18 |
huggingface-router | openai-chat | https://router.huggingface.co/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-18 |
amazon-bedrock | openai-chat | Required apiURL | bearer | tools, stream, modes function, tiers standard/flex/priority | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-17 |
azure-foundry | openai-chat | Required apiURL | api-key-header | tools, stream, modes function, tiers standard/priority | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-17 |
vertex-ai | openai-chat | Required apiURL | bearer | tools, stream, modes function | exact google/gemma-4-26b-a4b-it-maas: thinking, no native structured output, structured modes json_object -> function, default thinking max; prefix google/gemini-, gemini-: structured modes native -> function -> json_object | source 1, source 2, source 3 | 2026-08-18 |
databricks | openai-chat | Required apiURL | bearer | tools, stream, modes function, tiers standard/priority | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-08-17 |
baseten | openai-chat | https://inference.baseten.co/v1 | bearer | tools, stream, structured, modes native -> function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
groq | openai-chat | https://api.groq.com/openai/v1 | bearer | tools, stream, structured, modes native -> function, tiers standard/flex/priority | exact openai/gpt-oss-20b, openai/gpt-oss-120b: thinking, default thinking max; exact qwen/qwen3.6-27b: thinking, default thinking max | source 1, source 2, source 3 | 2026-08-18 |
cerebras | openai-chat | https://api.cerebras.ai/v1 | bearer | tools, stream, structured, modes native -> function, tiers standard/flex/priority | exact gpt-oss-120b: thinking, default thinking max; exact gemma-4-31b: thinking, default thinking max | source 1, source 2, source 3 | 2026-08-18 |
deepinfra | openai-chat | https://api.deepinfra.com/v1/openai | bearer | tools, stream, modes function, tiers standard/priority | prefix deepseek-ai/DeepSeek-R1: thinking, default thinking max | source 1, source 2, source 3 | 2026-08-18 |
sambanova | openai-chat | https://api.sambanova.ai/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
nebius | openai-chat | https://api.tokenfactory.nebius.com/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
novita | openai-chat | https://api.novita.ai/v3/openai | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
hyperbolic | openai-chat | https://api.hyperbolic.xyz/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
siliconflow | openai-chat | https://api.siliconflow.com/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
friendli | openai-chat | https://api.friendli.ai/serverless/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
cloudflare-workers-ai | openai-chat | Required apiURL | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
featherless | openai-chat | https://api.featherless.ai/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
nscale | openai-chat | Required apiURL | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
ovhcloud | openai-chat | Required apiURL | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
scaleway | openai-chat | https://api.scaleway.ai/v1 | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
nvidia-nim | openai-chat | Required apiURL | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
runpod-vllm | openai-chat | Required apiURL | bearer | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
sagemaker-vllm | openai-chat | Required apiURL | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
vllm | openai-chat | http://localhost:8000/v1 | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
ollama | openai-chat | http://localhost:11434/v1 | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
lm-studio | openai-chat | http://localhost:1234/v1 | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
llama-cpp | openai-chat | http://localhost:8080/v1 | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
localai | openai-chat | http://localhost:8080/v1 | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
baseten-engine | openai-chat | Required apiURL | bearer (optional) | tools, stream, modes function | Conservative defaults; use modelInfo for model-specific features | source | 2026-08-17 |
typesafe | typesafe-system-one | https://api.typesafe.ai | bearer | structured, modes native | Conservative defaults; use modelInfo for model-specific features | source 1, source 2 | 2026-09-15 |