博文

Nano Banana Pro API: 3 Model IDs Compared, from $0.03 per Image

If you see three model ids for Nano Banana Pro (Google's Gemini 3 Pro Image) on an API gateway, the short answer is: use gemini-3-pro-image by default, keep gemini-3-pro-image-preview only if it is already in your code and you want $0.068 instead of $0.08 at 1K and 2K, and send work that can wait to gemini-3-pro-image-gemini , a flat $0.03 per image that is much slower with reference images. Plan for 30 to 90 seconds per image. The rest of this post backs each point with numbers. Disclosure up front: the numbers below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. Latency and failure figures are from our production traffic over the last 7 days (sample size not disclosed), rounded to whole seconds and whole percentages. Three ids, one model: which to use All three ids are Nano Banana Pro with the same request body and response shape. T...

GPT Image 2 API: Which Model ID to Use, from $0.01 per Image

If you are calling GPT Image 2 from code, the short version is this: use gpt-image-2 for most work, at a flat $0.025, $0.03 or $0.05 per image for 1K, 2K or 4K; use gpt-image-2-all when you need low or high quality, or the synchronous response shape of the OpenAI SDK; use gpt-image-2-lite at $0.01 per image for drafts. Plan for about a minute per 2K image, poll for up to ten minutes rather than three, and read the real width and height off every result. The rest of this post backs each of those numbers. Disclosure up front: the numbers below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. All three GPT Image 2 model ids above run on it. Latency and failure figures are measured on our production traffic over the last 7 days (sample size not disclosed), rounded to whole seconds and whole percentages. Three model ids, one model All three ids...

Seedream 5.0 API: Flash vs Pro, from $0.03 per Image

Picking a Seedream 5.0 tier for an API integration? The short version: use Flash for volume work, transparent PNG edits and splitting an image into layers, at a flat $0.03 per image and about 16 seconds for a 2K render; use Pro when an edit must stay inside a marked region or the image carries a lot of text, at $0.03 per image at 1K and $0.06 at 2K, and plan for about a minute per image. Neither tier renders 4K. Disclosure up front: the numbers below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. Both Seedream 5.0 tiers from ByteDance are on it, as doubao-seedream-5-0-flash and doubao-seedream-5-0-pro . Latency and failure figures are from our production traffic over the last 7 days (sample size not disclosed), rounded. Flash or Pro: the short answer Both tiers use the same endpoint and request fields, so switching is a one-string change...

Veo 3.1 API: Which Model ID to Use, from $0.08 per 8-Second Clip

For most Veo 3.1 work through an API, the short version is this: use veo-3.1-fast-fhd by default ($0.088 for an 8-second 1080p clip with audio), drop to veo-3.1-fast-hd ($0.08, 720p) when someone is waiting and 720p is enough, and keep veo-3.1 , Google's standard tier, for the final cut at $1.00 a clip. Plan for a median of about three and a half minutes per 1080p Fast clip, and poll or use a webhook instead of holding a connection open. The rest of this post backs each of those points with numbers. Disclosure up front: the numbers below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. Timings and failure rates are measured on our production traffic over the last 7 days (sample size not disclosed) and rounded; prices are what our billing charges as of October 27, 2026. Which Veo 3.1 id for which job All three ids share one endpoint and ...

Seedance 2.0 Fast API: 480p and 720p Video from $0.071 per Second

If you are putting Seedance 2.0 Fast into a product, here is the short version. It is the 480p and 720p tier of ByteDance's Seedance 2.0 and takes the same inputs as the full model. A finished clip takes close to three minutes, so build around a task id and a webhook, not a blocking call. Billing is per second of output, from $0.071 at 480p and $0.153 at 720p. The rest of this post puts numbers behind each point. Disclosure up front: the measurements below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. Latency, failure and output figures come from our production traffic over the last 7 days (sample size not disclosed), rounded to whole seconds and whole percent. Fast, standard or Mini All three Seedance 2.0 tiers take the same request body, so moving a job from one to another means changing one string. seedance-2.0-fast renders 480p...

Wan 3.0 Video API: 30-Second Clips from $0.045 per Second

If you want Alibaba's Wan 3.0 video model behind an API, the short version is this: set resolution on every request, run batches on the standard tier and keep the prime tier for when a person is waiting, plan for about four minutes per clip of 10 seconds or less and about 16 minutes for longer shots, and expect per-second prices from $0.045 at 480P. The rest of this post backs each point with numbers. Disclosure up front: the numbers below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. Latency and failure figures are from our production traffic over the last 7 days (sample size not disclosed), standard tier only, rounded to whole seconds and percentages. What Wan 3.0 does, and which tier to pick Wan 3.0 is one model name, wan-3.0-video , that covers four kinds of job: Text to video : a prompt and nothing else. First frame, or first...

Kling 4.0 API: 30-Second AI Video with Sound, from $0.157 per Second

If you want Kling 4.0 in production, the short version is: call it as kling-4.0 on an asynchronous video endpoint, expect a 720p clip to take about three and a half minutes at the median and about ten minutes at p90, and plan on $0.157 per second at 720p or $0.235 at 1080p. Pick an older Kling id when you need silent output, a fixed price per clip or automatic rerouting. The rest of this post backs up each number. Disclosure up front: the numbers below come from apimodels.app, which is the API I work on. apimodels.app is a multi-model API gateway: one API key and one OpenAI-compatible endpoint for image, video, audio and language models. Kuaishou's Kling video models are part of that catalog. Measured figures are from our production traffic over the last 7 days (sample size not disclosed), October 18 to 25, 2026, rounded to whole seconds. Which Kling model id to use Four Kling video ids share one endpoint, POST /v1/video/generations , and one key. They differ in length, inpu...