OfoxAI es un gateway de IA unificado que te da acceso a 100+ modelos — GPT, Claude, Gemini, DeepSeek, Kimi, Qwen, GLM y más — con una sola API key y un único saldo. Habla los protocolos OpenAI, Anthropic y Gemini, factura por uso sin cuota de plataforma y publica un SLA del 99.9% sobre la capa de plataforma.

¿Cómo uso la API de OfoxAI?

Sustituye tu base URL actual de OpenAI, Anthropic o Google por el endpoint equivalente de OfoxAI y usa tu API key de OfoxAI. Para solicitudes compatibles con OpenAI es https://api.ofox.io/v1. Tu código existente sigue funcionando sin cambiar de SDK. La documentación completa está en https://ofox.io/docs.

¿Qué modelos ofrece OfoxAI?

Más de 100 modelos de texto de las familias GPT, Claude, Gemini, DeepSeek, Kimi, Qwen, GLM, Grok, MiniMax y Doubao, además de modelos multimedia: Seedance, Wan y HappyHorse para vídeo, y GPT-Image, Nano Banana y Seedream para imágenes. El catálogo publica el precio actualizado y las capacidades de cada uno.

¿Funciona OfoxAI en todo el mundo?

Sí. OfoxAI opera una red de aceleración global con nodos edge desplegados en varias regiones, lo que ofrece a los desarrolladores de cualquier parte del mundo un acceso a la API fiable y de baja latencia.

¿Qué herramientas de programación con IA funcionan con OfoxAI?

Claude Code, Codex, OpenCode, Zed, Cline, Chatbox, CherryStudio, OpenClaw, Kilo y Aider funcionan — sirve cualquier herramienta que acepte un base URL de API personalizado. Cambia el base URL y la API key; no hace falta tocar el código.

OfoxAI3 Minutos a Todos los LLMs

100+ modelos, una sola API key

Qwen, DeepSeek, GLM y Kimi desde $0.44/M — además de GPT, Claude y Gemini. Cambia una URL base y conserva tu SDK.

100+

Modelos Top

99.9%

Uptime

~300ms

Latencia

Obtener API Key Explorar Modelos

Seedance 2.5: hasta un 20% de descuento720p desde $0.24/s · 1080p $0.48/s ~~$0.60~~. Clips de hasta 30 s con audio sincronizado.

Modelos de código

Razonamiento, refactorización y agentes con una API.

Modelos de imagen

Genera, edita y perfecciona contenido visual.

Modelos de vídeo

Crea vídeo desde texto o imágenes, también con audio.

Partner autorizado de proveedores cloud líderes — 100+ modelos, una API

OpenAI

Claude

Gemini

DeepSeek

Qwen

Kimi

Doubao

GLM

Mistral

OpenAI

Claude

Gemini

DeepSeek

Qwen

Kimi

Doubao

GLM

Mistral

Generación de imagen y vídeo

No solo texto: también imagen y vídeo

Cada pieza es una generación real del modelo que aparece en ella: imágenes desde $0.03/imagen y vídeo desde $0.04/s.

Doubao Seedream 5.0 ProAn emperor penguin chick looking up in a blizzard, down feathers ruffled by the wind, a blurred adult colony behind, cold palette, hyper-real wildlife photography, 16:9 horizontal composition, no text, no watermark, no logo anywhere in the image

Seedance 2.5One single continuous take, no cuts, no edits. A cluttered 1980s suburban basement with faux wood-panel walls, retro posters and exposed ceiling joists. Four teenage friends argue passionately around a circular wooden table lit by one hanging shaded lamp: a boy in a mustard yellow raglan shirt throws his head back in defeat, a lanky boy in an oversized plaid flannel and suspenders presides behind a cardboard screen, a wild-haired boy in a faded denim vest buries his face in his palms, and a round-faced boy in a maroon cable-knit sweater and oversized tortoiseshell glasses thrusts both hands upward. Colourful miniatures and multi-sided dice sit untouched on the felt. The camera pushes in slowly from a low angle. Warm tungsten key, amber and wood-brown palette, shallow depth of field, realistic cinematic footage, 16:9, no text anywhere in frame.

Nano Banana ProAn abandoned greenhouse reclaimed by ferns, broken glass panes filtering late afternoon light into visible beams, damp air, photorealistic wide shot, no text, no watermark, no logo anywhere in the image

Nano Banana 2A market stall of spices in open sacks, deep reds and ochres, hands of a vendor reaching in from the left, warm shaded daylight, no text, no watermark, no logo anywhere in the image

Wan 2.7One single continuous take, no cuts, no edits. A quiet bedroom at night. In the middle of the room stands a free-standing doorway whose frame glows with slowly shifting rainbow LED light; through the opening, instead of the wall behind it, sunlit drifting clouds and open sky are visible. The camera drifts slowly forward toward the doorway. Soft ambient darkness, saturated rainbow spill across the bedding and walls, cinematic, shallow depth of field, 16:9, no text anywhere in frame.

GPT Image 2A weathered brass diving helmet resting on wet dock planks at dawn, sea mist behind, cold blue light with one warm lamp reflection, photorealistic, shallow depth of field, no text, no watermark, no logo anywhere in the image

Qwen-Image 3.0A winter night in an old lane, red lanterns glowing through falling snow, bicycle tracks pressed into the snow leading deep into the lane, warm and cool light meeting, 16:9 horizontal composition, no text, no watermark, no logo anywhere in the image

Seedance 2.0One single continuous take, no cuts, no edits. A packed football stadium erupts at night as a last-minute goal goes in. From pitch level the camera rises past a front-row supporter leaping up with both arms raised and a scarf whipping overhead, confetti drifting through the floodlight beams while the crowd blurs into motion behind him. Slight handheld shake, hyper-realistic, cinematic floodlight, shallow depth of field, 16:9, no readable text, no logos anywhere in frame.

Doubao Seedream 5.0 LiteClose-up of a vintage typewriter, light catching the metal body, paper curling slightly from the roller, a warm desk lamp from the right, wooden desktop, shallow depth of field, 16:9 horizontal composition, no text, no watermark, no logo anywhere in the image

Qwen-Image 3.0 ProA sea of clouds over granite peaks before a storm, a windswept pine branch in the foreground, distant ridges receding layer by layer into mist, cool cyan palette, cinematic telephoto compression, 16:9 horizontal composition, no text, no watermark, no logo anywhere in the image

Seedance 2.0 FastOne single continuous take, no cuts, no edits. A vast dark wet beach at twilight. A thin sheet of water creeps slowly across black sand and turns it into a perfect shimmering mirror. A lone distant human silhouette stands motionless while its reflection wobbles gently with the water surface. On the horizon a faint band of soft pink light pulses against a deep navy sky. The camera holds completely still. Ultra-detailed water physics, serene and melancholic, 16:9, no text anywhere in frame.

GPT Image 1.5A stack of worn hardcover books beside folded reading glasses, warm lamp light from the right, dark background, still life photography, no text, no watermark, no logo anywhere in the image

Nano Banana 2 LiteA hammock strung between two palms over white sand, calm turquoise water behind, bright noon light, calm and minimal, no text, no watermark, no logo anywhere in the image

Wan 2.6One single continuous take, no cuts, no edits. A cold late-15th-century stone courtroom lit by iron candle sconces. In the first row of wooden benches sits a pale young woman with ice-blue eyes and blonde hair, wearing a dark charcoal wool dress, a brown hooded cloak and an antique brooch. Beyond her an accused woman with dark wavy chestnut hair, olive skin and a faded rust-red shawl stands before a heavy table where a stern inquisitor and two jurors are seated. The camera moves in one slow handheld arc past the seated woman toward the tribunal. Candle flicker, shallow depth of field, deep shadows, documentary realism, 16:9, no text anywhere in frame.

MAI Image 2.5 ProA cluttered watchmaker's bench seen from above, tweezers and gears scattered around a half-assembled movement, warm task lamp light, macro realism, no text, no watermark, no logo anywhere in the image

Doubao Seedream 4.5A stream in an autumn forest, fallen leaves drifting on the water, sunlight through the canopy scattering into pools of light, a long exposure giving the water a silky texture, 16:9 horizontal composition, no text, no watermark, no logo anywhere in the image

Seedance 2.0 MiniA stylish silent brand promotional film for pink sneakers, cut fast with a new shot roughly every half second. Opens on a dramatic hero shot of a pair of vivid pink sneakers, then breaks into multiple dancers wearing them: breakdance footwork, spins, freezes and windmills, constant movement in every shot, performers bursting toward the lens. Motion trails, afterimages, ghosting and motion blur, whip pans, orbit moves, crash zooms and rapid tracking. Clean fashion-editorial framing, Swiss-inspired graphic composition, luxury sportswear commercial grade. Ends on a massive iconic hero shot of the pink sneakers. 16:9, no text, no lettering, no logos anywhere in frame.

MAI Image 2.5A ceramic teapot on a linen cloth beside a window, steam rising, soft overcast daylight, muted earth tones, still life photography, no text, no watermark, no logo anywhere in the image

Ver modelos de imagen Ver modelos de vídeo

API multimodal: texto, imagen y vídeo

Texto, imagen y vídeo — una sola clave

Cambia el endpoint y el ID del modelo: la clave, el saldo y el panel siguen iguales. Imágenes desde $0.03/imagen, vídeo desde $0.04/s.

POST /v1/videos{  "model": "bytedance/seedance-2.0",  "prompt": "El zorro avanza y las flores se mecen",  "duration": 6, "aspect_ratio": "4:3",  "frame_images": [{ "type": "image_url",    "image_url": { "url": "…/fox-01.jpg" },    "frame_type": "first_frame" }]}

DespuésGET /v1/videos/{id} → mirror_urls[0]

Fotograma de ejemplo de Seedance 2.0: El zorro avanza y las flores se mecen

clip de 6 sSeedance 2.0desde $0.07/sdemo predefinida

Seedance 2.0 Minidesde $0.04/s Seedance 2.5desde $0.11/s Seedance 2.0 Fastdesde $0.06/s Seedance 2.0desde $0.07/s Wan 2.7desde $0.10/s

Todos los modelos de vídeoTodos los modelos

Precios de API de LLM

Modelos frontier desde $0.44/M

Precios en vivo de nuestro catálogo, actualizados automáticamente.

MODELOPROVEEDORCAPACIDADESCONTEXTOPRECIO / UNIDAD

OPEN WEIGHTModelos insignia y económicos de DeepSeek, Qwen, Zhipu y Moonshot

DeepSeek V4 Flash Vision ExpVisiónFunctionsRazonamientoCaché+11M$0.44/$1.32/M

Qwen3.8 MaxVisiónFunctionsRazonamientoCaché+3+21M$2/$6/M

GLM-5.3-10%FunctionsRazonamientoCachéWeb+11Mdesde$1.26$1.4/$3.96$4.4/M

Kimi K3VisiónFunctionsRazonamientoCaché+2+11M$3/$15/M

FRONTIERGPT, Claude, Gemini y Grok

GPT-5.6 Sol-50%VisiónFunctionsRazonamientoCaché+2+11M$2.5$5/$15$30/M

Claude Fable 5VisiónFunctionsRazonamientoCaché+2+11M$10/$50/M

Gemini 3.7 Flash-50%VisiónFunctionsRazonamientoCaché+5+41M$0.75$1.5/$3.75$7.5/M

Grok 4.6FunctionsRazonamientoCaché500K$2/$6/M

GPT Image 2VisiónCaché—$5/$30/M

Seedance 2.5up to -20%Video—desde $0.11/s

Ver todos los modelos Mejor precio · hasta -80%

Ranking de uso de modelos y herramientas

Lo que los desarrolladores realmente ejecutan aquí

Cuota de tokens de toda la plataforma durante las últimas 13 semanas: las mismas cifras que nuestro leaderboard público, no una lista escogida a mano.

Uso anonimizado y agregado

MODELOSRanking completo

#MODELOTENDENCIA 13 SEMCUOTA