Seven new models, text-to-video and web search support, deeper usage visibility
New model lineup: Claude Fable 5.1, Gemini 3.8 Flash, and five more
New
Seven new models land on Gateway this week across two new deployment vendors.
- Claude Fable 5.1 by Anthropic
- Gemini 3.8 Flash by Google, with zero data retention
- DeepSeek V4 Pro 0813, MiniMax M3, and Qwen3.8 Max through Fireworks AI
- GLM-5.3 Flash through Baseten, Particle AI, and Wafer, with zero data retention
- Wafer joins as a new model vendor
Text-to-video generation through fal.ai
New
Gateway now routes text-to-video generation through fal.ai, including models such as Seedance 2.5. Video generation sits behind the same request path as every other model on Gateway.
Interfaze web search engine
New
Select Interfaze as a web search engine on Gateway, backed by the OpenWebSearch API. It sits alongside Gateway's other search engines for any request that needs live web results.
OpenAI's tool_search on /v1/openai/responses
New
Gateway adds support for OpenAI's hosted tool_search tool on /v1/openai/responses, supporting the deferred tool-loading round trip that Codex relies on.
Usage and Spend gains a By Routing Policy view, Logs gains a rollup strip
New
Usage and Spend adds a By Routing Policy tab, with a detail page showing spend, token rollups, and the requests routed through each policy. Filtered Logs results now show a rollup summary strip above them, covering models routed, prompt and completion tokens, cache reads and writes, and estimated cost.