AIREITER2026-09-09 — GPT-Image 2.5 launched for image generation and editing
GPT-Image 2.5 is now on AIReiter as two models on one page: Flare for fast everyday stills, and Sunburst when an edit needs to keep the rest of the frame stable. Generate from a prompt or attach reference images, then pick 1K, 2K, or 4K and one of fourteen aspect options.
- Text-to-image and reference-guided editing with up to ten input images
- 1K, 2K, and 4K output, plus API-only custom pixel sizes
- Flare and Sunburst on the same playground at /image/gpt-image-2-5
2026-09-02 — Chat prices aligned to official list rates for GPT, Grok, Claude, and Gemini
AIReiter updated token prices on GPT, Grok, Claude, and Gemini chat models. Each model page still shows the live rate next to the official list price; the table on the page is the source of truth.
- GPT and Gemini chat models are billed at 30% of the official list price (70% off).
- Grok chat models are billed at 20% of the official list price (80% off).
- Claude chat models are billed at 50% of the official list price (50% off). Cache-read and cache-write follow the same ratio where those meters apply.
View pricing · GPT-5.6 Sol · Grok 4.6 · Claude Sonnet 5 · Gemini 3.7 Flash
2026-08-31 — Gemini Omni 1.1 Flash Ext launched for 4K credit video
Gemini Omni 1.1 Flash Ext is now available on AIReiter for 4–10 second clips at 720p, 1080p, or 4K, billed in credits from text, stills, or a reference clip.
- Length is limited to 4, 6, 8, or 10 seconds. 720p and 1080p share one price row; 4K is a higher row.
- Start from a prompt, up to seven stills, one clip up to 10 seconds, or stills plus that clip.
- A reference video uses a flat per-generation price instead of the duration dropdown.
Open Gemini Omni 1.1 Flash Ext
2026-08-28 — Flow Music launched for generate, edit, stems, and clip video
Flow Music is now available in the AIReiter audio generator for prompt-to-track generation and follow-on editing on the same clip.
- Generate a track from a sound prompt, then extend, replace a section, cover, or split stems without leaving the audio tool.
- Lyrics, upload, download, and a short video clip from an existing track are included as follow-on actions.
- Each action uses its own model ID on the API; the playground keeps them on one Flow Music surface.
2026-08-28 — Gemini Omni Flash Ext launched for 1080p and 4K credit video
Gemini Omni Flash Ext is now available on AIReiter as a credit-billed video SKU for 4–10 second clips at 720p, 1080p, or 4K.
- Length is limited to 4, 6, 8, or 10 seconds. 720p and 1080p share one price row; 4K is a higher row.
- Start from a prompt, a first-frame still, one or three reference images, or one short source clip.
- First Frame cannot include a reference video. A reference clip uses a flat per-generation row instead of the duration dropdown.
2026-08-28 — Kling v3 Omni launched for native-audio video generation
Kling v3 Omni is now available on AIReiter for text, multi-image reference, first/last frames, and reference-video workflows, with optional native audio.
- Charged per output second at 720p, 1080p, or 4K. Native audio raises the 720p/1080p rate; 4K is the same with or without audio.
- A reference video uses the video-input rate. Feature length follows Video Length; base follows the uploaded clip length.
- Playground fields map to the same kling_v3_omni parameters on the API.
2026-08-27 — Z-Image Turbo launched for fast photorealistic stills
Z-Image Turbo is now available on AIReiter as a low-cost text-to-image model for photorealistic output with English and Chinese text rendering.
- One flat rate per image, with nine aspect ratios and no resolution tier or reference-image surcharge.
- Each request returns one image. An optional seed reproduces a similar composition from the same prompt.
- The playground and API expose prompt, aspect ratio, and seed only.
2026-08-27 — GLM-5.3 Flash launched for multimodal flash-tier chat
GLM-5.3 Flash is now available in the AIReiter chat playground with image input, a large context window, streaming, and flash-tier token pricing.
- Send images alongside text in the same request. GLM-5.3 Flash is natively multimodal.
- Keeps a 1M-token context window and 128K output for high-volume chat, extraction, and agent inner loops.
- Current input, output, and cache-read rates are shown on the model page.
2026-08-26 — Seed Audio 1.0 launched for prompt-to-audio
Seed Audio 1.0 is now available in the AIReiter audio generator for text-to-audio clips billed per output second.
- Describe the sound, voice, or bed you need and choose an output format in the audio tool.
- Charged per completed second, with a maximum of 120 seconds per request. Failed tasks are not billed.
- Use it for short voice beds, sound design, and prompt-driven audio rather than full song editing.
2026-08-26 — Suno launched for music generation, extend, cover, and lyrics
Suno is now available in the AIReiter audio generator for full-track generation and follow-on editing of an uploaded or generated clip.
- Generate a song from a prompt, with optional custom title/style and instrumental mode.
- Extend or cover an existing clip, or request lyrics as a separate low-cost action.
- Playground actions map to suno, suno_extend, suno_cover, and suno_lyrics on the API.
2026-08-25 — Wan 3.0 launched for 2–30s video from text, frames, files, and pages
Wan 3.0 is now available on AIReiter for text, first/last frames, mixed references, and document or webpage to video.
- Output 2–30 seconds at 480p, 720p, or 1080p. Native audio is on by default and does not change the per-second price.
- Attach images, clips, or audio as references. Reference-video seconds are added to generated seconds.
- File/Webpage mode turns a document or public page into a clip on the same model ID.
2026-08-24 — Qwen3.8 Max launched for 1M-context chat and tools
Qwen3.8 Max is now available in the AIReiter chat playground with a 1M-token context window, PDF chat, and billed built-in tools.
- OpenAI-compatible Chat Completions and Responses, with streaming in the playground.
- Optional web search, image search, and related tools are priced per call on the model page.
- Current input, output, cache-read, and cache-write rates are shown above the playground.
2026-08-21 — HappyHorse 1.1 launched for 720p and 1080p video
HappyHorse 1.1 is now available on AIReiter as a per-second video model at 720p and 1080p.
- Charged per generated second at the selected resolution.
- Use the 1.1 page when you want the newer HappyHorse SKU rather than the 1.0 generate-and-edit page.
- Playground fields map to the same happyhorse_1_1 parameters on the API.
2026-08-21 — HappyHorse 1.0 launched for generate and edit video
HappyHorse 1.0 is now available on AIReiter for text-to-video, image-to-video, multi-reference, and source-video editing on one page.
- Generation workflows support 720p or 1080p output and whole-second durations from 3 to 15 seconds.
- Video editing bills source duration plus generated output duration. Other modes bill generated duration only.
- Switch among the four workflows without changing the happyhorse_1_0 model ID.
2026-08-19 — Grok Imagine Image 2.0 launched for generation and reference editing
Grok Imagine Image 2.0 is now available on AIReiter for prompt-driven stills and reference-guided edits, with an Ext mode on the same page for higher output counts.
- Primary mode: 1K or 2K, Low or Medium quality, up to three reference images, and 1–10 outputs per request.
- Ext mode on the same page is a flat per-image SKU for 1–12 outputs, without the primary quality and resolution grid.
- Each input reference on the primary model adds a one-time per-request surcharge, not a per-output surcharge.
2026-08-14 — GLM-5.3 launched for flagship chat
GLM-5.3 is now available in the AIReiter chat playground with streaming, a large context window, and token pricing.
- Use it for production chat, long documents, and bilingual Chinese/English work.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-14 — Gemini 3.7 Flash launched for fast chat
Gemini 3.7 Flash is now available in the AIReiter chat playground for responsive assistants, extraction, and high-volume text workloads.
- Streaming output with live token usage after completion.
- Current input, output, and cache-read rates are shown above the playground.
- Use it as the newest Gemini Flash SKU on AIReiter.
2026-08-14 — Grok 4.5 launched for chat and API
Grok 4.5 is now available in the AIReiter chat playground with streaming and token pricing, including cache-read rates.
- Use the Grok 4.5 page when an integration is already pinned to this model ID.
- Current input, output, and cache-read rates are shown above the playground.
- The same model ID is used in the playground and the chat API.
2026-08-14 — Grok 4.6 launched for chat and API
Grok 4.6 is now available in the AIReiter chat playground with streaming and token pricing, including cache-read rates.
- Use 4.6 as the current Grok chat SKU for new integrations.
- Current input, output, and cache-read rates are shown above the playground.
- The same model ID is used in the playground and the chat API.
2026-08-13 — Midjourney V8.1 launched with generate, retexture, and remove-bg
Midjourney V8.1 is now available on AIReiter for image generation. Retexture and background removal run as follow-on actions on the same family page.
- Each generate request returns up to four images. Credits are charged per request, not per surviving image after content review.
- Draft/Fast and Standard/HD controls are exposed on the V8.1 page.
- Remove-bg returns one image per request. Retexture returns up to four.
2026-08-13 — Midjourney V7 launched with generate, retexture, and remove-bg
Midjourney V7 is now available on AIReiter for image generation. Retexture and background removal run as follow-on actions on the same family page.
- Each generate request returns up to four images. Credits are charged per request, not per surviving image after content review.
- Draft, Fast, and Turbo speeds are exposed on the V7 page.
- Remove-bg returns one image per request. Retexture returns up to four.
2026-08-12 — Seedream 5.0 Pro Layerize launched for per-layer image splits
Seedream 5.0 Pro Layerize is now available on AIReiter to split a generated result into billed layers after completion.
- Billed per generated layer, with 2–17 layers per request. 1K uses a lower per-layer rate than Auto / 1.5K / 2K.
- A minimum balance is required to start. The final charge follows the completed layer count.
- Use Seedream 5.0 Pro when you want generation or editing rather than a layered export.
Open Seedream 5.0 Pro Layerize
2026-08-11 — Krea 2 Turbo launched for 1K and 2K stills
Krea 2 Turbo is now available on AIReiter as a fast image model with 1K and 2K resolution SKUs.
- Choose 1K for drafts and 2K for higher-detail stills. Price follows the selected resolution.
- The playground and API share the same krea_2_turbo model ID.
- The model page is live with SEO metadata and pricing aligned to the production configuration.
2026-08-06 — FLUX 3 Video launched for text, keyframes, and source-video extension
FLUX 3 Video is now available on AIReiter for text-to-video, ordered keyframes, and source-video extension with native audio.
- HD and FHD length SKUs cover 5–20 seconds.
- A reference-video (v2v) path uses a higher price tier than text or image starts.
- Playground fields map to the same flux_3_video parameters on the API.
2026-08-06 — Qwen Image 3.0 Pro launched for 2K stills and references
Qwen Image 3.0 Pro is now available on AIReiter for text-to-image and reference-guided stills, with a higher 2K rate than standard 3.0.
- 1K and 2K output, with 1–6 images per request.
- Attached reference images add a small per-request surcharge.
- Use Pro when 2K quality is the priority.
2026-08-06 — Qwen Image 3.0 launched for 1K and 2K stills
Qwen Image 3.0 is now available on AIReiter for text-to-image and reference-guided stills at a flat 1K/2K per-image rate.
- 1K and 2K share the same per-image rate. Each request can return 1–6 images.
- Attached reference images add a small per-request surcharge.
- Use 3.0 for volume; use 3.0 Pro when 2K quality is the priority.
2026-08-04 — MiniMax M3 launched for chat and API
MiniMax M3 is now available in the AIReiter chat playground with streaming and token pricing.
- Use it for production chat and high-volume text workloads.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-04 — GLM 5.2 launched for production chat
GLM 5.2 is now available in the AIReiter chat playground with streaming and token pricing.
- Use it when an integration is already standardized on the GLM 5.2 model ID.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-04 — Doubao Seed 2.1 Turbo launched for chat and API
Doubao Seed 2.1 Turbo is now available in the AIReiter chat playground with streaming and token pricing.
- Positioned for fast, cost-conscious chat and extraction.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-04 — Kimi K2.7 Code launched for coding chat
Kimi K2.7 Code is now available in the AIReiter chat playground as a coding-oriented SKU.
- Use it for implementation, debugging, and code-centric chat.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-04 — DeepSeek V4 Flash launched for high-volume chat
DeepSeek V4 Flash is now available in the AIReiter chat playground as the cheaper V4 option for high-volume text.
- Use Flash when request volume matters more than peak reasoning depth.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-04 — DeepSeek V4 Pro launched for stronger V4 chat
DeepSeek V4 Pro is now available in the AIReiter chat playground as the stronger DeepSeek V4 SKU.
- Use Pro when the task needs more depth than V4 Flash.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-08-03 — MiniMax H3 launched for quality-tier video generation
MiniMax H3 is now available on AIReiter for video generation with quality-tier per-second billing and mixed references.
- Generated duration and reference-video duration are billed together at the selected quality rate.
- The first five reference images are free; each additional image adds a per-image surcharge.
- Playground fields map to the same minimax_h3 parameters on the API.
2026-07-27 — Claude Opus 5 launched for flagship chat
Claude Opus 5 is now available in the AIReiter chat playground for complex reasoning, demanding coding work, and multi-step analysis.
- Streaming output with live token usage after completion.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-24 — Kimi K3 launched for long-context chat
Kimi K3 is now available in the AIReiter chat playground for large working sets of code, documents, research, and agent history.
- Use it when context is the bottleneck and one request needs to carry a large working set.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-23 — Claude Fable 5 launched for long-form writing
Claude Fable 5 is now available in the AIReiter chat playground for detailed writing, careful synthesis, and high-quality long-form work.
- Use it when structure, voice, and revision quality matter more than a quick first draft.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-23 — Claude Sonnet 5 launched for production chat
Claude Sonnet 5 is now available in the AIReiter chat playground for balanced coding, analysis, writing, and production assistants.
- Use it as a practical default when every request does not need the Opus tier.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-23 — Claude Opus 4.7 launched for chat and API
Claude Opus 4.7 is now available in the AIReiter chat playground for agentic coding, debugging, and professional analysis.
- Use it when an integration is pinned to the Opus 4.7 model ID.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-23 — Claude Opus 4.8 launched for chat and API
Claude Opus 4.8 is now available in the AIReiter chat playground for agentic coding, difficult debugging, and professional knowledge work.
- Use it when an integration is pinned to the Opus 4.8 model ID.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-23 — Gemini 3.5 Flash Lite launched for cost-sensitive chat
Gemini 3.5 Flash Lite is now available in the AIReiter chat playground for fast, cost-sensitive text processing, extraction, and lightweight chat.
- Use it for short, frequent requests where latency and request economics matter most.
- Current token rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-23 — Gemini 3.6 Flash launched for fast chat
Gemini 3.6 Flash is now available in the AIReiter chat playground for responsive assistants, rapid content processing, and high-volume API workflows.
- Streaming output with visible token usage after completion.
- Current input, output, and cache-read rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-22 — FLUX.2 Pro launched for 1K and 2K stills
FLUX.2 Pro is now available on AIReiter as an image model with 1K and 2K resolution SKUs.
- Price follows the selected resolution.
- The playground and API share the same flux_2_pro model ID.
- The model page is live with SEO metadata and pricing aligned to the production configuration.
2026-07-14 — Kling 3.0 Motion Control launched for clip-driven motion
Kling 3.0 Motion Control is now available on AIReiter to drive a character or subject from an uploaded clip.
- Billed by uploaded duration and resolution at 720p or 1080p.
- Use it when source motion should be preserved while the visual subject changes.
- Playground fields map to the same kling_3_0_motion_control parameters on the API.
2026-07-14 — Kling 3.0 Turbo launched for 720p and 1080p video
Kling 3.0 Turbo is now available on AIReiter for 3–15 second clips at 720p or 1080p.
- Length and resolution determine the SKU price.
- Use Turbo when you want the faster Kling 3.0 path rather than Omni or Motion Control.
- Playground fields map to the same kling_3_0_turbo parameters on the API.
2026-07-10 — GPT-5.6 Sol launched for flagship GPT-5.6 chat
GPT-5.6 Sol is now available in the AIReiter chat playground for the most demanding coding, reasoning, and agent workflows in the GPT-5.6 family.
- Use Sol for the hardest requests; keep routine traffic on Terra or Luna.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-10 — GPT-5.6 Terra launched for production GPT-5.6 chat
GPT-5.6 Terra is now available in the AIReiter chat playground as the balanced production SKU in the GPT-5.6 family.
- Use Terra as a practical default for coding, analysis, writing, and automation.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-10 — GPT-5.6 Luna launched for fast, economical GPT-5.6 chat
GPT-5.6 Luna is now available in the AIReiter chat playground for fast, economical chat, extraction, classification, and high-volume text automation.
- Use Luna for frequent or high-volume requests that do not need Sol or Terra depth.
- Current input, output, cache-read, and cache-write rates are shown on the model page.
- The same model ID is used in the playground and the chat API.
2026-07-09 — Seedream 5.0 Pro launched for high-precision image generation
Seedream 5.0 Pro is now available on AIReiter as a ByteDance image model for higher-precision generation, image editing, and multi-reference creative workflows.
- The new seedream_v5_pro model supports text-to-image, image-to-image, and image editing workflows.
- Users can work with up to 10 reference images and choose 1K or 2K output resolution.
- The model page is live with SEO metadata, OpenAPI projection, and pricing specs aligned to the production model configuration.
2026-07-07 — Seedance 2.5 launched for ByteDance video generation
Seedance 2.5 is now available on AIReiter for text-to-video, image-to-video, and reference-video workflows, including speech/audio references and first/last-frame stills.
- Generate at 480p or 720p. With a reference video, input seconds plus output seconds are billed together at the lower combined rate.
- Reference capacity includes a large image set plus multiple video and audio files, with optional generated audio and MP4 or MOV output.
- Playground controls match the seedance2_5 API parameters.
Open Seedance 2.5 · View Seedance 2.5 API docs
2026-07-05 17:13:07 — Super-resolution collection expanded with dedicated upscaling models
AIReiter expanded the super-resolution collection with practical upscalers for photos, product shots, AI artwork, compressed images, and print-oriented assets.
- PrunaAI P-Image-Upscale is positioned as the fast default for general cleanup and target-megapixel output.
- Google Upscaler adds simple 2x and 4x image enlargement for straightforward upscaling tasks.
- Recraft Creative Upscale and Recraft Crisp Upscale cover stronger texture enhancement and sharper presentation-ready outputs.
Open Super Resolution collection · PrunaAI P-Image-Upscale · Google Upscaler · Recraft Creative Upscale · Recraft Crisp Upscale
2026-07-06 — Product updates page added for model and tool release notes
AIReiter now has a dedicated updates page for human-written release notes covering new tools, model page changes, API documentation updates, and platform improvements.
- Release notes are maintained as reviewed content instead of being generated from provider metadata.
- The first version is English-only to avoid duplicate localized pages before translations are ready.
- Entries focus on AIReiter product changes rather than upstream provider pricing or hidden supplier details.
2026-06-29 — GPT Image 2 trial pages added for 1K drafts and product mockups
Two GPT Image 2 landing flows were added to help users test lower-cost 1K drafts and product mockup prompts before moving into higher-resolution final work.
- The 1K generator page is positioned for fast draft exploration before paid 2K or 4K usage.
- The product mockup page targets ecommerce, ads, and product concept validation workflows.
- Both pages use dedicated preview assets and structured SEO metadata.
Open GPT Image 2 1K generator · Open GPT Image 2 product mockup generator
2026-06-29 — Lyric video generator landing page added for music-video search intent
A dedicated lyric video generator page was added to separate music-video use cases from the general AI video generator surface.
- The page targets creators who need lyric clips, music visuals, and social-ready video variations.
- It keeps the production workflow connected to the main AI video generation experience.
- The route has its own SEO head, FAQ schema, and localized navigation shell support.