Wan 2.2 text-to-image — fast tier, async via /v1/jobs.
Alibaba Cloud Model Studio exposes Wan text-to-image through its own image-generation task API and model-specific DashScope / Bailian conventions.
TheRouter presents the same served model behind the routed id wan/wan2.2-t2i-flash on its OpenAI-compatible image generation surface, with catalogue-level parameters limited to prompt, size, n, and seed.
wan2.2-t2i-flash is the fast tier of Alibaba's Wan 2.2 text-to-image model. It generates images from text prompts at $0.04 per image — the right default for high-volume creative work, thumbnails, social-post variants, and exploratory iteration.
Image generation runs asynchronously: submit a job to `/v1/jobs` and receive an `img_`-prefixed job ID, then poll until the job completes and fetch the resulting URL. This pattern follows OpenAI's async job convention — your existing SDK helpers should work with a model-ID swap.
| Type | Rate |
|---|---|
| Image | $0.027 / image |
| Per image | $0.027 / image |
Per-image flat fee. Alibaba's own published USD price: wan2.2-t2i-flash $0.025/image (alibabacloud.com/help/en/model-studio/model-pricing, International pane, read 2026-07-29).
| Model family | Wan 2.2 / Tongyi Wanxiangwanxiang.aliyun.com ↗ | verified |
| Generation mode | Text-to-imagealibabacloud.com/help/en/model-studio ↗ | verified |
| Tier positioning | Flash / fast tieralibabacloud.com/help/en/model-studio ↗ | verified |
| Supported request controls | prompt, size, n, seed | verified |
| Training cutoff / architecture | Not publicly disclosed for this served tier | unknown |
| Benchmark | Distribution | Score | Source |
|---|---|---|---|
Public quality benchmark Alibaba's public Model Studio pages document the served text-to-image capability and pricing tier, but do not publish a comparable named quality benchmark for wan2.2-t2i-flash. | — | Not publicly disclosed | — |
Public latency benchmark The flash label indicates fast-tier positioning, but no public p50/p95 latency result for TheRouter's live route is available in the cited operator sources. | — | Not publicly disclosed | — |
Public prompt-adherence benchmark No named prompt-adherence score was found in the free, cited sources for this no-budget curation pass; keep the gap explicit instead of inventing a score. | — | Not publicly disclosed | — |
Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.
Recommended: use the async API
Image generation typically takes 30–180s, beyond the edge sync timeout. The examples below use the ?async=true submit + poll pattern. Read the full async image generation & edit guide →
# 1) Submit job (returns 202 immediately with a polling URL).
# Image generation takes 30-180s — always use the async path in production.
JOB=$(curl -s -X POST "https://api.therouter.ai/v1/images/generations?async=true" -H "Content-Type: application/json" -H "Authorization: Bearer $THE_ROUTER_API_KEY" -d '{
"model": "wan/wan2.2-t2i-flash",
"prompt": "A cinematic product render with soft studio lighting"
}' | python3 -c "import sys,json;print(json.load(sys.stdin)['id'])")
echo "submitted: $JOB"
# 2) Poll until terminal (succeeded / failed / cancelled / expired).
while :; do
R=$(curl -s "https://api.therouter.ai/v1/jobs/$JOB" -H "Authorization: Bearer $THE_ROUTER_API_KEY")
S=$(echo "$R" | python3 -c "import sys,json;print(json.load(sys.stdin)['status'])")
echo "status: $S"
case "$S" in
succeeded) echo "$R" | python3 -c "import sys,json;print(json.load(sys.stdin)['unsigned_urls'][0])"; break ;;
failed|cancelled|expired) echo "$R"; exit 1 ;;
esac
sleep 5
doneThe catalogue exposes prompt, size, n, and seed. Use seed for reproducible exploration when the upstream honors it.
curl https://api.therouter.ai/v1/images/generations -H "Authorization: Bearer $THEROUTER_API_KEY" -H "Content-Type: application/json" -d '{
"model": "wan/wan2.2-t2i-flash",
"prompt": "Three thumbnail concepts for a productivity app launch, clean vector style",
"size": "1024x1024",
"n": 3,
"seed": 42
}'TheRouter's generated model catalogue lists wan/wan2.2-t2i-flash as the lower-cost Wan 2.2 text-to-image tier, alongside wan/wan2.2-t2i-plus for higher-fidelity output. Operators should document the snippet-verification gap until a funded live API run is approved.
Use flash for exploration, bulk variants, and lower-cost drafts. Use plus when the image is a final deliverable, a brand asset, or a paid campaign visual where fidelity matters more than unit cost.
No. TheRouter lists this route as text-to-image with text input and image output. For image-conditioned editing or image-to-image workflows, choose an image-editing model instead of this flash T2I route.
No. This no-budget curator run did not execute paid snippet verification. Treat the examples as shape guidance until a human operator approves a funded verification run.
| source | URL | retrieved | |
|---|---|---|---|
| Model family | wanxiang.aliyun.com ↗ | 2026-08-05 | verified |
| Generation mode | alibabacloud.com/help/en/model-studio ↗ | 2026-08-05 | verified |
| Tier positioning | alibabacloud.com/help/en/model-studio ↗ | 2026-08-05 | verified |
| Supported request controls | — | — | verified |
| Training cutoff / architecture | — | — | unknown |
| Public quality benchmark | alibabacloud.com/help/en/model-studio ↗ | 2026-08-05 | unknown |
| Public latency benchmark | alibabacloud.com/help/en/model-studio ↗ | 2026-08-05 | unknown |
| Public prompt-adherence benchmark | wanxiang.aliyun.com ↗ | 2026-08-05 | unknown |
| Wan 2.2 T2I flash and plus tiers appear in TheRouter catalogue | alibabacloud.com/help/en/model-studio ↗ | 2026-08-05 | verified |