Tasks and inputs
MiniMax H3 Image Highspeed is listed for Image-to-image, Text-to-image. It accepts text, image and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Nov 17, 2026
MiniMax H3 Image Highspeed is the faster hosted tier of MiniMax's H3 Image generator and editor, at double the image-token price of the standard tier.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative USD per 1M tokens: $2.30 text input, $3.00 image input, $7.60 image output; about $0.035 per 1K text-to-image request. Launch promo 50% off outputs >=16.6MP until 2026-11-17 16:00 UTC. Confirm on your own account.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Speed is the whole product here. The standard H3 Image tier is already quick; MiniMax positions it at 1.5–2 times the pace of mainstream alternatives. Highspeed is pitched at 3–4 times. You pay for the difference on every image.
This is a serving tier, not a new model release. It went live alongside MiniMax H3's image model on 10 November 2026 (the launch blog is dated the 11th) and is called as MiniMax-H3-Image-Highspeed through the MiniMax API.
It has no separate weights. The open-weight model card never mentions a highspeed mode, so treat this as hosted-only. Some model listings flag it as open-weight; we found nothing to support that.
Everything the standard tier can do should carry over: text-to-image, reference-guided generation and editing, up to 16 reference images, and output up to roughly 16 megapixels. MiniMax has not published Highspeed's own resolution limits or any quality trade-off, so verify at the sizes you need.
| Per 1M tokens (indicative, USD) | Highspeed | Standard |
|---|---|---|
| Text input | $2.30 | $2.30 |
| Image input | $3.00 | $1.50 |
| Image output | $7.60 | $3.80 |
Text tokens cost the same. Image tokens cost exactly double.
MiniMax's worked examples make it concrete. A 1K text-to-image call is about $0.035 on Highspeed versus $0.022 on standard. An edit with three reference images at 1K is about $0.060 versus $0.036. At 4K the list price is about $0.130 per image, halved under a launch promotion for outputs of at least 16.6 million pixels that ends on 17 November 2026 at 16:00 UTC.
Prompt expansion adds text tokens on top. Check your own account for current figures.
Think about where a person is waiting.
If images are generated in a queue nobody watches, the standard tier is the rational choice. Half the image cost, same model.
Latency claims need your own numbers. Fire the same thirty prompts at both tiers from the region where your servers run. Log submit-to-URL time at the 50th and 95th percentile, and cost per image. Then look at a handful of outputs side by side at full size; if Highspeed achieves its speed with any quality compromise, that is where it will show.
No independent quality metrics existed at launch, so your comparison is the only one that counts. Other MiniMax models, including MiniMax image-01, are on the MiniMax provider page.
Compare the model on the work you need to complete.
MiniMax H3 Image Highspeed is listed for Image-to-image, Text-to-image. It accepts text, image and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
The listed route has starting rates of $3.00 input and $7.60 output per million tokens. Choose the output size and quality your design needs. Include drafts and revisions when budgeting an accepted image.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
recraft
Recraft V4.1 Flash V2 is Recraft's fastest, cheapest image model: 1K raster output in about 1.1 s for $0.007, quality close to Recraft V4.1.
MiniMax
MiniMax H3 Image is an open-weight 33.1B image generation and editing model producing native ~16MP output in six steps, with up to 16 reference images.
OpenAI
GPT-Image-2.6 Flare is OpenAI's fast image generation and editing model for interactive iteration, with SVG, animated GIF and markup-guided edits.
OpenAI
GPT-Image-2.6 Sunburst is OpenAI's detail-first image generation and editing model, adding SVG, animated GIF output and markup-guided edits.
krea
Krea 3 is Krea's in-house image model for text-to-image and instruction-based edits at 1K or native 2K, priced from $0.03 per image.
Alibaba
Qwen-Image-2.1-Turbo creates and edits images, including reference-guided and transparency workflows; hosted billing is separate from local deployment.