August 2026 added eight production models to Poyo.ai, covering image, video, and chat in one month. The same month also upgraded Suno Music into a single workspace and expanded Grok Imagine Video 1.5 with additional workflows and 1080p output.
This monthly update covers only models that are live on Poyo.ai today. Prices below match the public model pages as of September 2, 2026. Confirm the current rate on the model page before you ship a production workflow.
How to choose
Start with the job, then open the matching model page.
| If you need | Start with | Why |
|---|---|---|
| Lower-cost video, 2–30 seconds | Wan 3.0 | Text, image, and reference workflows at 480p, 720p, or 1080p |
| Faster Wan 3.0 turnaround | Wan 3.0 Video Prime | Same workflow family, priced for speed |
| Longer clips and many references | Seedance 2.5 | 4–30 seconds and up to 50 reference files |
| First/last frames, extension, keyframes, native audio | FLUX 3 | Five video workflows in one family |
| Dense layouts and multilingual text in images | Qwen Image 3 | Long prompts, fine text, 1K and 2K output |
| Text-to-image plus light reference editing | Grok Imagine Image 2.0 | Up to 3 input images and 4 outputs per request |
| Long-running agents, coding, and knowledge work | Grok 4.6 | Chat API at 20% below official token rates |
| Multimodal agents, tools, and long context | Gemini 3.7 Flash | Chat API with cached-input pricing |
| Songs, sound effects, voices, and mashups | Suno Music | 24 music tools in one workspace |
Video: the largest August change
Seedance 2.5
Seedance 2.5 went live on August 8 for longer, reference-heavy video. Use it when a clip needs more than a prompt: first and last frames, images, video, and audio can all travel in one request.
- Model ID:
seedance-2.5 - Duration: 4–30 seconds
- Output: 480p, 720p, or 1080p
- References: up to 30 images, 10 videos, and 10 audio files, with 50 files maximum per request
- Optional synchronized audio through
generate_audio
| Output | With reference video | Without reference video |
|---|---|---|
| 480p | 17 credits / $0.085 per billed second | 28 credits / $0.14 per second |
| 720p | 38 credits / $0.19 per billed second | 63 credits / $0.315 per second |
| 1080p | 68.5 credits / $0.3425 per billed second | 114 credits / $0.57 per second |
When a reference video is present, billed duration equals output duration plus reference-video duration. See the Seedance 2.5 API docs for request fields. For a side-by-side with MiniMax H3, read MiniMax H3 vs Seedance 2.5.
Wan 3.0 and Wan 3.0 Video Prime
Wan 3.0 and Wan 3.0 Video Prime went live on August 25. Both support text-to-video, image-to-video, and multimodal reference-to-video. Choose standard Wan 3.0 when cost matters more; choose Prime when generation time matters more.
Shared capabilities:
- 2–30 second output at 480p, 720p, or 1080p
- One or two images as the first frame and optional last frame
- Up to 10 reference images, 5 reference videos, 5 reference audio clips, and one document or public webpage
- Optional generated audio on all three workflows
| Resolution | Wan 3.0 | Wan 3.0 Video Prime |
|---|---|---|
| 480p | 10 credits / $0.05 per second | 13.6 credits / $0.068 per second |
| 720p | 20 credits / $0.10 per second | 28 credits / $0.14 per second |
| 1080p | 40 credits / $0.20 per second | 56 credits / $0.28 per second |
Model IDs:
- Wan 3.0:
wan3.0-text-to-video,wan3.0-image-to-video,wan3.0-reference-to-video - Wan 3.0 Video Prime:
wan3.0-prime-text-to-video,wan3.0-prime-image-to-video,wan3.0-prime-reference-to-video
FLUX 3
FLUX 3 became a production video family on August 6, with five workflows instead of a single text-to-video endpoint.
flux-3/text-to-videoflux-3/image-to-videoflux-3/first-last-frame-to-videoflux-3/extend-videoflux-3/keyframes-to-video
Standard workflows are 34 credits / $0.17 per second at 720p and 58 credits / $0.29 per second at 1080p. Video extension is 82 credits / $0.41 per second at 720p and 106 credits / $0.53 per second at 1080p. Native audio is part of the family.
Use FLUX 3 when you need shot control, continuation, or placed keyframes rather than a one-shot generation. Deeper walkthroughs: FLUX 3: Everything You Need to Know and FLUX 3 vs FLUX 2.
Grok Imagine Video 1.5 upgrade
This is an upgrade, not a new catalog entry. Grok Imagine Video 1.5 now covers text-to-video, single-frame image-to-video, and reference-to-video with 1–7 images. Output includes 480p, 720p, and 1080p.
| Item | Price |
|---|---|
| 480p | 14.5 credits / $0.0725 per second |
| 720p | 25 credits / $0.125 per second |
| 1080p text or image mode | 45 credits / $0.225 per second |
| Input image | 2 credits / $0.01 per image |
Image: two complementary options
Qwen Image 3 and Qwen Image 3 Pro
Qwen Image 3 moved from a preview page to a live generator on August 6. It is the better fit for long prompts, information-dense layouts, small text, and multilingual graphics.
- Model IDs:
qwen-image-3,qwen-image-3-pro - Output: 1K or 2K
- Tasks: text-to-image and image-to-image
| Variant | 1K | 2K |
|---|---|---|
| Qwen Image 3 | 4.8 credits / $0.024 | 4.8 credits / $0.024 |
| Qwen Image 3 Pro | 6.4 credits / $0.032 | 12 credits / $0.06 |
Input images add 0.5 credits / $0.0025 each. Full background: Qwen Image 3.0: Everything You Need to Know.
Grok Imagine Image 2.0
Grok Imagine Image 2.0 went live on August 15. One model ID handles text-to-image and editing with up to three input images.
- Model ID:
grok-imagine-image-2.0 - Resolution: 1K or 2K
- Quality: low or medium
- Up to 4 images per request
- Output: JPEG, PNG, or WebP
| Quality / resolution | Price per image |
|---|---|
| Low / 1K | 8 credits / $0.04 |
| Medium / 1K | 12 credits / $0.06 |
| Low / 2K | 12 credits / $0.06 |
| Medium / 2K | 16 credits / $0.08 |
| Input image add-on | 2 credits / $0.01 per input image |
Chat: two new production models
Grok 4.6
Grok 4.6 went live on August 15 for long-running agents, coding, engineering, and knowledge work. Call it through the OpenAI-compatible Chat Completions API.
- Model ID:
grok-4.6 - Input: 320 credits / $1.60 per 1M tokens (official $2.00)
- Output, including reasoning tokens: 960 credits / $4.80 per 1M tokens (official $6.00)
That is 20% below the official token rates. Docs: OpenAI-compatible chat.
Gemini 3.7 Flash
Gemini 3.7 Flash also went live on August 15. Use it for agentic coding, multimodal reasoning, tool use, and long-context work when you want Google's GA Flash model rather than a larger flagship.
- Model ID:
gemini-3.7-flash - Input: 120 credits / $0.60 per 1M tokens (official $0.75)
- Cached input: 12 credits / $0.06 per 1M tokens (official $0.075)
- Output, including thinking tokens: 600 credits / $3.00 per 1M tokens (official $3.75)
Docs: Gemini native format.
Music: Suno workspace upgrade
On August 28, Suno Music became one workspace with 24 tools. You can switch between generation, editing, mashups, reusable voices, stem splitting, and format conversion without leaving the page. Each tool shows its model ID, price, and API docs.
New and upgraded capabilities include:
- Generate Mashup: blend two audio tracks with style, vocal, model version, and duration controls. 12 credits / $0.06 per generation.
- Generate Sounds: sound effects and ambience with loop, tempo, and key controls. 2.5 credits / $0.0125 per generation.
- Suno Voice: validate a voice, create a reusable Voice ID, regenerate the verification phrase, and check availability. Voice operations are free.
- Generate MIDI: 1 credit / $0.005 per generation.
- Music generation and extension now support V5, V5.5, Style Persona, Voice Persona, and duration up to 360 seconds.
- Editing covers, extensions, section replacement, added vocals, added instrumentals, vocal separation, and 12-stem splits.
API docs: Generate Music.
Platform changes you can use now
These are the August product changes that affect how you work on Poyo.ai:
- A dashboard overview that keeps history and usage in one place.
- Generation history that shows result previews again.
- A clearer model market, with live discount pricing on model cards when a promotion is active.
- New developer entry points: MCP, CLI, Skill, plus n8n and ComfyUI guides.
- Vietnamese and Indonesian site locales.
August 2026 model cheat sheet
| Live date | Model | Type | Model IDs | Page |
|---|---|---|---|---|
| Aug 6 | Qwen Image 3 / Pro | Image | qwen-image-3, qwen-image-3-pro | /models/qwen-image-3 |
| Aug 6 | FLUX 3 | Video | flux-3/text-to-video and four related endpoints | /models/flux-3 |
| Aug 8 | Seedance 2.5 | Video | seedance-2.5 | /models/seedance-2-5 |
| Aug 15 | Grok 4.6 | Chat | grok-4.6 | /models/grok-4-6 |
| Aug 15 | Gemini 3.7 Flash | Chat | gemini-3.7-flash | /models/gemini-3-7-flash |
| Aug 15 | Grok Imagine Image 2.0 | Image | grok-imagine-image-2.0 | /models/grok-imagine-image-2 |
| Aug 25 | Wan 3.0 | Video | wan3.0-text-to-video, wan3.0-image-to-video, wan3.0-reference-to-video | /models/wan-3-0 |
| Aug 25 | Wan 3.0 Video Prime | Video | wan3.0-prime-text-to-video, wan3.0-prime-image-to-video, wan3.0-prime-reference-to-video | /models/wan-3-0-video-prime |
August upgrades, not counted in the eight launches: Grok Imagine Video 1.5 and Suno Music.
How to start
- Open the model page and run one test generation with the live tool.
- Copy the model ID into your existing Poyo workflow: async submit and status polling for media, or Chat Completions / Gemini native format for chat.
- If you are wiring this into an internal tool, start from MCP, CLI, n8n, or ComfyUI.
After August: early September launches
Two more models went live just after this August update window:
- Gemini Omni 1.1 Flash on September 1, for text-to-video, first and last frames, reference images, and reference-video workflows from 360p through 4K. Guide: Gemini Omni 1.1 Flash API.
- Claude Fable 5.1 on September 2, for long-horizon coding and knowledge work with a 1M context window.
They are not part of the August eight. Use the links above if you are already evaluating the newest video and chat options.