v1.1.14

[1.1.14] — 2026-07-05

OpenAI-compatible provider fidelity release (part two). The fifth provider-readiness remediation phase aligns the xAI, OpenRouter, Moonshot, NVIDIA NIM, Perplexity, and Novita providers on request/response correctness, and adds shared improvements that benefit every OpenAI-compatible provider. No breaking API changes relative to v1.1.13.

Fixed

• Reasoning & cache token usage (xAI, OpenRouter, and every OpenAI-compatible provider): usage reported in the nested OpenAI form (prompt_tokens_details.cached_tokens, completion_tokens_details.reasoning_tokens) is now decoded into the reported token counts on both the chat and streaming paths instead of being dropped.
• Perplexity Sonar metadata: citations, search results, images, and related questions are now surfaced (under a provider_metadata response field) instead of being discarded.
• Perplexity model discovery: the always-failing discovery call (it targeted a nonexistent endpoint and returned the wrong model family) is removed; the static Sonar model list is authoritative.
• Perplexity model list: the retired sonar-reasoning model is removed (sonar-reasoning-pro replaces it).
• NVIDIA NIM embeddings: now route through the shared embeddings path, so errors surface consistently and input_type is forwarded.
• xAI image responses: decode into the canonical image response type and route errors through the shared error envelope; the user field is now forwarded, and size/quality/style are reported as dropped when a Grok image model ignores them.
• parallel_tool_calls: a client-supplied parallel_tool_calls is now forwarded to the provider instead of being dropped at the HTTP boundary.

Changed

• Live model discovery added for Moonshot and NVIDIA NIM (self-updating model lists); the Moonshot and xAI static fallback lists were refreshed.
• Base-URL validation is now enforced at construction for xAI, OpenRouter, Moonshot, NVIDIA NIM, Perplexity, and Novita.
• OpenRouter embeddings: OpenRouter now supports the embeddings endpoint via the shared helper.

Contributors

Thanks to everyone who shipped this release:

• v1.1.14 - OpenAI-Compat Providers II (#324) — @MitulShah1

---

Full changelog: https://github.com/ferro-labs/ai-gateway/blob/v1.1.14/CHANGELOG.md