Documentation
Models
A short model model id is the only string you need. The gateway picks the channel.
Every Orux AI model is reachable by a stable short model id such as claude-opus-4.7, gpt-5.5, deepseek-v3.1. Model ids are versioned (claude-opus-4.7 ≠ claude-opus-46) and never silently change behaviour. Use GET /api/v1/models for a live, per-account list.
GET
/api/v1/modelsBearer sk-app-…Catalogue at a glance#
38
Chat
9
Image
17
Video
4
Audio
1
Embedding
1
Rerank
Model id convention#
Model ids are short, lowercase, dash-separated. No modality prefix — claude-opus-4.7, gpt-5.5, deepseek-v3.1, veo-3.1-quality, kling-3.0, suno-v5, gpt-image-2 are all canonical.
Fallback chains#
A bare model id maps to multiple upstream channels with priorities. If the primary times out, errors with 5xx, or trips its circuit breaker, traffic shifts to the next channel. For async tasks (POST /api/v1/tasks), the GET response body carries a `channel` field telling you which upstream finally served the job.
Live model list
The catalogue below is fetched live from the gateway — what you see is what your keys can actually call.
Full model matrix#
Search any model id, name or description. Click View for the per-model parameter sheet under the corresponding modality.
| Model ID | Model | Modality | Spec | Capabilities | Top params | Doc |
|---|---|---|---|---|---|---|
gpt-5-2 | GPT-5.2 OpenAI flagship; large-context reasoning and tool use. | chat | 400K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-5-pro | GPT-5 Pro Highest-tier OpenAI reasoning model, deeper chain-of-thought, longer answers. | chat | 400K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-5-codex | GPT-5 Codex Code-specialised GPT-5; better instruction following on programming tasks. | chat | 400K ctx | ToolsStream | messagestemperaturetop_p | View → |
gpt-codex | GPT Codex Legacy GPT code model retained for some callers. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
claude-opus-4-5 | Claude Opus 4.5 Anthropic top-tier model: best at long-form reasoning, code review and agentic tool use. Supports prompt caching. | chat | 200K ctx | ToolsVisionCacheStream | messagestemperaturetop_p | View → |
claude-sonnet-4-5 | Claude Sonnet 4.5 Balanced Claude tier — fast, cheaper, still tool/vision capable. | chat | 200K ctx | ToolsVisionCacheStream | messagestemperaturetop_p | View → |
claude-haiku-4-5 | Claude Haiku 4.5 Smallest Claude — sub-second latency, good for chatbots and routing. | chat | 200K ctx | ToolsCacheStream | messagestemperaturetop_p | View → |
gemini-3-pro | Gemini 3 Pro Google flagship; 2M context, native multimodal. | chat | 2000K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gemini-3-flash | Gemini 3 Flash Fast tier of Gemini 3. | chat | 1000K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
grok-3 | Grok 3 xAI conversational model with web tools. | chat | 256K ctx | ToolsStream | messagestemperaturetop_p | View → |
grok-3-mini | Grok 3 Mini Smaller, cheaper Grok 3. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
deepseek-ai/DeepSeek-V3.1 | DeepSeek V3.1 DeepSeek general-purpose flagship. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
deepseek-ai/DeepSeek-R1 | DeepSeek R1 DeepSeek reasoning model; emits chain-of-thought. | chat | 64K ctx | Stream | messagestemperaturetop_p | View → |
zai-org/glm-4.5 | GLM-4.5 Zhipu GLM 4.5. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
zai-org/glm-4.6 | GLM-4.6 Zhipu GLM 4.6. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
zai-org/glm-4.7 | GLM-4.7 Zhipu GLM 4.7, latest tier. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
zai-org/glm-4.5v | GLM-4.5V GLM vision-language variant. | chat | 32K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
Qwen/Qwen3-max | Qwen3 Max Alibaba Qwen3 max. | chat | 256K ctx | ToolsStream | messagestemperaturetop_p | View → |
Qwen/Qwen3.5-plus | Qwen 3.5 Plus Qwen 3.5 plus tier. | chat | 256K ctx | ToolsStream | messagestemperaturetop_p | View → |
Qwen/Qwen3.6-max-preview | Qwen 3.6 Max (Preview) Qwen 3.6 max preview. | chat | 256K ctx | ToolsStream | messagestemperaturetop_p | View → |
ByteDance/doubao-seed-1.6 | Doubao Seed 1.6 ByteDance Doubao seed 1.6. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
ByteDance/doubao-2-0-mini | Doubao 2.0 Mini Doubao 2.0 mini. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
ByteDance/doubao-2-0-lite | Doubao 2.0 Lite Doubao 2.0 lite. | chat | 128K ctx | Stream | messagestemperaturetop_p | View → |
ByteDance/doubao-2-0-code-preview | Doubao 2.0 Code (Preview) Doubao 2.0 code-specialised preview. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
xiaomi/mimo-v2-flash | MiMo V2 Flash Xiaomi MiMo flash tier. | chat | 64K ctx | ToolsStream | messagestemperaturetop_p | View → |
xiaomi/mimo-v2-pro | MiMo V2 Pro Xiaomi MiMo pro tier. | chat | 128K ctx | ToolsStream | messagestemperaturetop_p | View → |
claude-opus-4-1-20250805 | Claude Opus 4.1 Previous-generation Claude Opus tier, date-pinned snapshot id. | chat | 200K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
claude-sonnet-4-5-20250929 | Claude Sonnet 4.5 Claude Sonnet 4.5, date-pinned snapshot id. | chat | 200K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
claude-haiku-4-5-20251001 | Claude Haiku 4.5 Claude Haiku 4.5, date-pinned snapshot id. | chat | 200K ctx | ToolsStream | messagestemperaturetop_p | View → |
gpt-5 | GPT-5 OpenAI GPT-5 base tier; 256K context. | chat | 256K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-5-pro-umv | GPT-5 Pro GPT-5 Pro variant with 256K context. | chat | 256K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-4o-mini | GPT-4o Mini Compact, low-cost GPT-4o mini. | chat | 128K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-4o | GPT-4o OpenAI GPT-4o multimodal model. | chat | 128K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-5-codex-umv | GPT-5 Codex Code-specialised GPT-5 variant with 256K context. | chat | 256K ctx | ToolsStream | messagestemperaturetop_p | View → |
gemini-3-pro-preview | Gemini 3 Pro (Preview) Gemini 3 Pro preview build; 2M context. | chat | 2000K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gemini-3-flash-preview | Gemini 3 Flash (Preview) Gemini 3 Flash preview build — fast tier. | chat | 1000K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gemini-2.5-pro | Gemini 2.5 Pro Gemini 2.5 Pro; 2M context, native multimodal. | chat | 2000K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gemini-2.5-flash-lite | Gemini 2.5 Flash Lite Lightweight, fast Gemini 2.5 Flash lite. | chat | 1000K ctx | ToolsVisionStream | messagestemperaturetop_p | View → |
gpt-image-2 | GPT-Image-2 OpenAI native image generation. Strong typography, clean compositions. | image | 1536x1536 | promptsizequality | View → | |
gemini-3-pro-image | Gemini 3 Pro Image Google Imagen variant via Gemini 3 Pro. Strong photo realism. | image | 1408x1408 | promptaspect_ration | View → | |
flux-kontext-pro | Flux Kontext Pro Black Forest Labs flux-kontext: prompt-driven edit & generation with reference images. | image | 2048x2048 | promptaspect_ratioprompt_upsampling | View → | |
flux-kontext-max | Flux Kontext Max Top-tier flux-kontext, higher quality, longer latency. | image | 2048x2048 | promptaspect_ratioreference_image_url | View → | |
runway-gen4-image | Runway Gen-4 Image Runway Gen-4 still-image generation. Up to 3 reference images. | image | 1920x1080 | promptaspect_ratioreference_images | View → | |
midjourney-v7 | Midjourney v7 Midjourney v7. Multi-step workflow: imagine → upscale / variation / reroll using returned task_id. | image | 2048x2048 | promptaspect_ratiostylize | View → | |
seedream-4-0 | Seedream 4.0 ByteDance Seedream 4.0 — strong Chinese-language prompts. | image | 4096x4096 | promptsizeaspect_ratio | View → | |
qwen-image-edit | Qwen Image Edit Alibaba Qwen image edit — region-targeted edits. | image | — | promptimage_url | View → | |
recraft-v3 | Recraft v3 Recraft v3 — vector-art friendly model. | image | — | promptstyle | View → | |
veo-3.1-quality | Veo 3.1 Quality Google Veo 3.1, quality tier. 8s fixed length. Accepts an input image as URL or base64 data URI. | video | ≤ 8s | Audio | promptaspect_ratioaudio | View → |
veo-3.1-fast | Veo 3.1 Fast Veo 3.1 fast tier — lower price, slightly less detail. | video | ≤ 8s | Audio | promptaspect_ratioaudio | View → |
sora2 | Sora 2 OpenAI Sora 2 — text-to-video, 5s or 10s. | video | ≤ 10s | promptdurationaspect_ratio | View → | |
runway-gen3-alpha-turbo | Runway Gen-3 Alpha Turbo Runway Gen-3 alpha turbo — image-to-video, 5/10s. | video | ≤ 10s | promptimage_urlduration | View → | |
runway-gen4 | Runway Gen-4 Runway Gen-4 — image- or text-to-video. | video | ≤ 10s | promptimage_urlduration | View → | |
runway-aleph | Runway Aleph Runway Aleph — edit existing video clips with text. | video | — | promptvideo_url | View → | |
kling-v21-master-i2v | Kling v2.1 Master (I2V) Kling 2.1 master image-to-video. | video | ≤ 10s | promptimage_urlduration | View → | |
kling-v21-master-t2v | Kling v2.1 Master (T2V) Kling 2.1 master text-to-video. | video | ≤ 10s | promptduration | View → | |
kling-v25-i2v-pro | Kling v2.5 Pro (I2V) Kling 2.5 pro image-to-video. | video | ≤ 10s | promptimage_urlduration | View → | |
kling-v25-t2v-pro | Kling v2.5 Pro (T2V) Kling 2.5 pro text-to-video. | video | ≤ 10s | promptduration | View → | |
kling-avatar-std | Kling Avatar (Std) Kling avatar standard — lip-sync animated avatar. | video | — | promptimage_urlaudio_url | View → | |
kling-avatar-pro | Kling Avatar (Pro) Kling avatar pro — higher fidelity. | video | — | promptimage_urlaudio_url | View → | |
hailuo-02 | MiniMax Hailuo 02 MiniMax Hailuo 02. Supports camera directives embedded in the prompt. | video | ≤ 10s | promptimage_urlduration | View → | |
grok-imagine-t2v | Grok Imagine (T2V) xAI Grok Imagine, text-to-video. | video | — | promptduration | View → | |
grok-imagine-i2v | Grok Imagine (I2V) xAI Grok Imagine, image-to-video. | video | — | promptimage_url | View → | |
topaz-upscale | Topaz Video Upscale Topaz video enhance / upscale to 4K. | video | — | video_urltarget_resolution | View → | |
infinitalk-audio | InfiniTalk (Audio-driven) Audio-driven talking avatar. Takes a driver audio and a still character image. | video | — | driver_audio_urldriver_image_url | View → | |
suno-v5 | Suno v5 Suno v5 music generation. Multi-step workflow: generate → extend / cover / lyrics / add-vocals / instrumental / mashup / stem / mv / wav / cover-img / replace / persona / upload-extend. | audio | — | promptlyricsinstrumental | View → | |
elevenlabs-v3 | ElevenLabs v3 (TTS) ElevenLabs v3 text-to-speech. | audio | — | textvoice_idmodel_id | View → | |
elevenlabs-music | ElevenLabs Music ElevenLabs music generation. | audio | — | promptduration | View → | |
runway-act-two | Runway Act-Two Runway Act-Two — drive a character video by a performance reference (audio + driver video). | audio | — | character_video_urldriver_video_url | View → | |
text-embedding-3-large | Text Embedding 3 Large Large embedding model for retrieval. | embedding | — | inputdimensions | View → | |
bge-reranker-v2-m3 | BGE Reranker v2 m3 Multilingual reranker. | rerank | — | querydocumentstop_n | View → |
70 models
Categories#
Chat & Reasoning
claude-opus-4.7, claude-sonnet-4.5, claude-haiku-4.5, gpt-5.5, gpt-5.4-mini, gpt-codex, gemini-3-pro, gemini-3-flash, deepseek-v3.1, deepseek-r1, qwen3-max, glm-46
Priced per million input/output tokens.
Image generation
gpt-image-2, gemini-3-pro-image, midjourney
Priced per image, by size and quality.
Video generation
veo-3.1-quality, veo-3.1-fast, sora-2-pro, kling-3.0, hailuo-02-i2v-pro, wan-26-i2v, seedance-1.5-pro, runway-aleph
Priced per second of output.
Audio & Music
suno-v5, suno-extend
Priced per song or per second.
List models programmatically#
List available models
curl https://orux.top/api/v1/models \
-H "Authorization: Bearer $ORUX_API_KEY"