Noryva: Powerful models. Create more. Spend less.

One account for your model APIs.

Models & pricing

Coding & text

GLM-5.3

Zhipu / Z.ai

Planned

Official reference price

Input
$1.40
Output
$4.40
USD · per 1M tokensDetails — billing conditions and source for GLM-5.3

Cached input: $0.26 per 1M tokens. Standard pay-as-you-go API prices; Coding Plan subscriptions are separate.

Reference checked . Planned for Noryva; availability unconfirmed.

Z.ai · international ↗ — GLM-5.3 official pricing

GLM-5.2

Zhipu / Z.ai

Planned

Official reference price

Input
$1.40
Output
$4.40
USD · per 1M tokensDetails — billing conditions and source for GLM-5.2

Cached input: $0.26 per 1M tokens. Standard pay-as-you-go API prices; Coding Plan subscriptions are separate.

Reference checked . Planned for Noryva; availability unconfirmed.

Z.ai · international ↗ — GLM-5.2 official pricing

Qwen3.8 Max

Alibaba Cloud

Planned

Official reference price

Input
$2.00
Output
$6.00
USD · per 1M tokensDetails — billing conditions and source for Qwen3.8 Max

Singapore / International deployment, up to 1M input tokens per request. Standard input excludes caching discounts. Other deployment regions have different rates.

Reference checked . Planned for Noryva; availability unconfirmed.

Alibaba Cloud · Singapore / International ↗ — Qwen3.8 Max official pricing

Qwen3.8 Flash

Alibaba Cloud

Planned

Official reference price

Input
$0.15
Output
$0.47
USD · per 1M tokensDetails — billing conditions and source for Qwen3.8 Flash

Singapore / International deployment, up to 1M input tokens per request. Standard input excludes caching discounts. Other deployment regions have different rates.

Reference checked . Planned for Noryva; availability unconfirmed.

Alibaba Cloud · Singapore / International ↗ — Qwen3.8 Flash official pricing

DeepSeek V4 Pro

DeepSeek

Planned

Official reference price

Input
$0.66–1.32
Output
$1.98–3.96
USD · per 1M tokensDetails — billing conditions and source for DeepSeek V4 Pro

Cached input: $0.022–0.044 per 1M tokens. Ranges show off-peak to peak pricing. Peak: Monday–Friday, 01:00–04:00 and 06:00–10:00 UTC; all other hours are off-peak. Uncached input shown above.

Reference checked . Planned for Noryva; availability unconfirmed.

DeepSeek · direct API ↗ — DeepSeek V4 Pro official pricing

DeepSeek V4 Flash

DeepSeek

Planned

Official reference price

Input
$0.22–0.44
Output
$0.66–1.32
USD · per 1M tokensDetails — billing conditions and source for DeepSeek V4 Flash

Cached input: $0.007–0.014 per 1M tokens. Ranges show off-peak to peak pricing. Peak: Monday–Friday, 01:00–04:00 and 06:00–10:00 UTC; all other hours are off-peak. Uncached input shown above.

Reference checked . Planned for Noryva; availability unconfirmed.

DeepSeek · direct API ↗ — DeepSeek V4 Flash official pricing

Hy4 preview

Tencent Hunyuan

Planned

Official reference price

Input
CNY 6.00
Output
CNY 18.00
per 1M tokensDetails — billing conditions and source for Hy4 preview

Cached input: CNY 0.30 per 1M tokens. Guangzhou online inference rate card. Preview model; prices are in Chinese yuan and have not been converted to USD.

Reference checked . Planned for Noryva; availability unconfirmed.

Tencent Cloud · Guangzhou ↗ — Hy4 preview official pricing

Video generation

Seedance 2.5

ByteDance

Planned

Official reference price

No video input
CNY 70
With video input
CNY 42
per 1M video tokensDetails — billing conditions and source for Seedance 2.5

480p / 720p online inference. 1080p list rates: CNY 77 / 46 per 1M video tokens (without / with video input). A 28% discount applies to 1080p until 17 Sep 2026, 14:00 UTC+8. Video input affects billable usage and minimum charges; these are not per-second prices.

Reference checked . Planned for Noryva; availability unconfirmed.

Volcengine · China ↗ — Seedance 2.5 official pricing

Seedance 2.0

ByteDance

Planned

Official reference price

No video input
CNY 46
With video input
CNY 28
per 1M video tokensDetails — billing conditions and source for Seedance 2.0

480p / 720p online inference. 1080p: CNY 51 / 31 per 1M video tokens; 4K: CNY 26 / 16 (without / with video input). Different resolutions consume different token quantities; a lower token rate does not imply a cheaper video. Video-input minimum charges apply.

Reference checked . Planned for Noryva; availability unconfirmed.

Volcengine · China ↗ — Seedance 2.0 official pricing

Wan 3.0

Alibaba Cloud

Planned

Official reference price

480p
$0.05
720p
$0.10
1080p
$0.20
USD · per billable second · list priceDetails — billing conditions and source for Wan 3.0

Singapore / International deployment. Billable seconds include input video plus successfully generated output. Standard list prices shown; the official page currently advertises a temporary 30% discount. Check the source for its current terms.

Reference checked . Planned for Noryva; availability unconfirmed.

Alibaba Cloud · Singapore / International ↗ — Wan 3.0 official pricing

Wan 3.0 Prime

Alibaba Cloud

Planned

Official reference price

480p
$0.068
720p
$0.14
1080p
$0.28
USD · per billable secondDetails — billing conditions and source for Wan 3.0 Prime

Singapore / International deployment. Billable seconds include input video plus successfully generated output.

Reference checked . Planned for Noryva; availability unconfirmed.

Alibaba Cloud · Singapore / International ↗ — Wan 3.0 Prime official pricing

Kling 3.0

Kuaishou / Kling

Version to confirm

Official reference price

720p
$0.084
1080p
$0.112
4K
$0.42
USD · per second · no native audioDetails — billing conditions and source for Kling 3.0

Reference configuration: Kling 3.0 without native audio. Audio and motion-control modes have different rates. The planned Noryva integration version is not yet confirmed. Developer API pricing is separate from consumer memberships.

Reference checked . Planned for Noryva; availability unconfirmed.

Kling · global developer API ↗ — Kling 3.0 official pricing

Coding and creative workflows.

Illustrative AI conversations. Hover to pause. Click or press Space to pause or resume. Use the left and right arrow keys to change examples.

NoryvaPlayground
Demo
You:

Only show active models in the picker.

lib/models.js

Filter the list before creating the options.

models.jsJavaScript
const options = models
  .filter(({ active }) => active)
  .map(({ id }) => ({ label: id, value: id }));

Inactive models are excluded.

Ask a follow-up…

Coding assistants & agents

Hover to pause. Tap or press Space to pause or resume. Swipe or use the left and right arrow keys to change films.

Image & video workflows

Cloud supply
at negotiated rates.

  • AWS
  • Google Cloud
  • Microsoft Azure
  • Volcengine
  • Tencent Cloud
  • Alibaba Cloud
Supply details

Our cloud supply partners include AWS, Google Cloud, Microsoft Azure, Volcengine, Tencent Cloud and Alibaba Cloud. Model availability, processing regions and retention terms depend on the selected service and route.

OpenAI-compatible text APIs.

curl "$NORYVA_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  --data @- <<JSON
{
    "model": "$NORYVA_MODEL",
    "messages": [{"role": "user", "content": "Write a Python function to deduplicate a list."}],
    "stream": true
}
JSON
Connection settings

Set NORYVA_BASE_URL to your console's API base URL, including /v1 without a trailing slash; NORYVA_API_KEY to your key; and NORYVA_MODEL to a supported text model ID.

Python and JavaScript use the OpenAI SDK. Image and video models have separate parameters and may use asynchronous tasks.

Start with an API key.