Telnyx / Service profile
Inference API · strict region AUS
Configuration dependentAn OpenAI-compatible API with a strict data-residency mode: requests sent to api.telnyx.com.au with region AUS run in that region or fail with HTTP 422.
Source review: 09 Nov 2026 · Service profile; exact model and processing route not established
Model coverage
Telnyx-hosted open models, e.g. moonshotai/Kimi-K2.6, GLM-5.3, DeepSeek V4.1 Flash, Gemma 4 26B A4B; AUS model list not published
API access
Self-serve account; OpenAI-compatible /v2/ai/chat/completions
Billing
Per 1M tokens, currency not labelled: e.g. Kimi K2.6 $0.665 / $4.00, GLM-5.3 $1.25 / $4.00, Gemma 4 26B $0.13 / $0.40
Capacity commitment
No reserved capacity required for this service category.
Storage
Chat Completions does not store request or response data. The Responses endpoint stores conversations under the account's data-locality setting.
Retention and training
Training use is governed by Telnyx terms and DPA; opt out through the account team.
Ownership and control
Telnyx LLC (US); Telnyx Australia Pty Ltd (ABN 79 623 603 725) is registered in Australia. Sydney GPUs were deployed in October 2025.
What to check before using it
- Send requests to api.telnyx.com.au with region AUS and mode strict.
- Probe the chosen model with a strict request; a 422 means it is not deployed in AUS.
- Use Chat Completions rather than Responses if you need no stored conversations, and review the DPA.
Still to verify
- Published list of models deployed in AUS
- Billing currency on the pricing page
- Ingress and logging locations
Sources and supported claims
Inference data residency ↗
In strict mode, inference is performed in the requested region or the request fails with 422; strict is not a blanket guarantee for ingress, logging or billing.
Checked 2026-11-09 · Scope: this service profileInference reference ↗
The region table maps AUS to Australia at api.telnyx.com.au and describes default routing as best-effort.
Checked 2026-11-09 · Scope: this service profileInference API pricing ↗
The pricing page lists per-1M-token model prices without a currency label.
Checked 2026-11-09 · Scope: this service profileSydney GPU release note ↗
Telnyx announced Sydney GPU deployment for voice AI and inference in October 2025.
Checked 2026-11-09 · Scope: this service profileAustralian Business Register ↗
Australian Business Register lists TELNYX AUSTRALIA PTY LTD, ABN 79 623 603 725.
Checked 2026-11-09 · Scope: this service profile