Skip to main content

Products

WaveAPI September update: use models from your coding assistant

Following the 2.0 release, WaveAPI added 16 models in late September, expanded coding assistant integrations, introduced Veo and Kling 4K options, and updated cache and multi-image billing.

2 min read
On this page
WaveAPI's Agent tools page
Agent tools: remote MCP, Agent Skill and AI docs. Screenshot captured on October 4, 2026, after the reporting period.

Following the WaveAPI 2.0 release, we added 16 models in late September: 12 text, three image and one video. You can also use WaveAPI from your coding assistant and select 4K output for Veo and Kling videos.

Connect your coding assistant

Text models gained Responses API support for clients such as Codex. Follow the Codex setup guide, or use the native Messages integration for Claude Code.

For image, video and audio tasks, connect remote MCP or install the Agent Skill. Model discovery, parameter lookup and cost estimates are free; submitting a generation task is billed. See the setup steps.

Choose 4K video output

Veo 3.1 and Kling v3 gained 4K options. SkyReels V4 and PixVerse V6 added video extension. For real-person references in Seedance, upload the media to the asset library and wait for approval before using it.

Veo 3.1 Fast configured for 4K output
Model page: 4K and eight seconds selected, with a quote shown before submission. No task was submitted; the apple is the page's sample output. Captured October 4, 2026. Check the quote when you make a call.

Check the updated billing rules

Cache hits on 13 OpenAI models are now billed at the cache-read rate. Multi-image requests are charged for the images actually returned, with excess reservations refunded. The pricing page also shows web search rates per 1,000 searches.

Web search rates for text models
Pricing page: web search rates for Qwen3.7-Flash and GPT-5 Nano. Captured October 4, 2026; prices and model counts reflect the page at that time.

Review existing integrations

  • Replace gpt-image-2.5 with gpt-image-2.5-flare.
  • Replace grok-imagine-1.0-video-rev with grok-imagine-video-1.5-rev. The Sora 2 models have been removed.
  • DeepSeek V4 time-based rates and some media model prices changed. Review the pricing changelog.

Website access is restricted in mainland China; API calls are outside this restriction. See the model changelog for the full record, or open Agent tools to get started.

View all