Model collections
Curated lists by use case and language: Spanish, Portuguese, Japanese, Chinese, Arabic, and more — plus coding, agents, video. All on one OpenAI-compatible API.
45 collections
Cheap AI models: low cost per token
Best cost-per-token for high volume, drafts, and classification without overpaying.
View collectionModels for testing and CI: fast and cheap
Fast and cheap for prompt validation, evals, and CI pipelines without burning budget.
View collectionBest AI models for coding and agents
For IDEs and agents: generation, refactors, debugging, and agentic workflows.
View collectionModels for AI apps: tools and structured outputs
For products: function calling, JSON schema, and agents that talk to your backend.
View collectionAI models for ads, copy, and creatives
Ad copy, hooks, landings, and UGC scripts — creative models plus cheap ones for volume.
View collectionAI video generation models
Text-to-video and image-to-video: Veo, Sora, Kling, Runway, and more on one gateway.
View collectionAI image generation models
Text-to-image for product, ads, and design: GPT Image, Flux, Ideogram, and more.
View collectionBest models for roleplay (RP) and creative writing
Character chat, RP, and narrative: models with strong style, context, and price.
View collectionEmbedding models for search and RAG
Vectors for RAG, semantic search, and clustering with OpenAI embeddings.
View collectionLong-context models: 200k to 1M tokens
200k–1M windows for repos, long PDFs, and agents with extended memory.
View collectionVision models: multimodal LLMs for images
Multimodal: understand images, visual docs, and charts in the same chat.
View collectionReasoning models: thinking and chain-of-thought
When the problem is hard: math, planning, deep debugging, and CoT.
View collectionModels with tool calling and function calling
Reliable function calling for agents, MCP, and workflows that hit your backend.
View collectionText-to-speech (TTS) models for voice
Voice from text: narration, IVR, assistants, and accessibility.
View collectionSpeech-to-text and transcription models
Audio-to-text transcription for meetings, calls, and captions.
View collectionBest models for autonomous AI agents
Autonomous agents: tools, planning, and long loops without losing the thread.
View collectionOpen-weight and open-source models via API
Llama, DeepSeek, Mistral, and more: open-weight power with API DX.
View collectionModels for RAG: answer over your knowledge
The generate side of RAG: cheap on the hot path, strong when retrieval is messy.
View collectionModels for chatbots and customer support
Support chatbots: low latency, consistent tone, controlled cost per conversation.
View collectionFast AI models: low latency
Priority: low TTFT. Autocomplete, routers, tip-as-you-type, and reactive UIs.
View collectionBest AI models for Spanish (LATAM)
Natural Spanish for Mexico and LATAM: support, copy, light legal, and local products.
View collectionBest AI models for Portuguese (Brazil)
Natural Brazilian Portuguese: CX, copy, and apps for the BR market.
View collectionBest AI models for English
Native product English: docs, US/UK support, and EN-first apps.
View collectionBest AI models for French
Natural French: Mistral + Claude/GPT/Gemini for FR and FR-CA.
View collectionBest AI models for German
Formal/informal German for DACH: CX, docs, and B2B copy.
View collectionBest AI models for Italian
Natural Italian for CX, ecommerce, and IT products.
View collectionBest AI models for Japanese
Japanese with proper keigo and tone for CX and JP apps.
View collectionBest AI models for Chinese (简体/繁體)
Simplified and Traditional Chinese: DeepSeek/Kimi + global flagships.
View collectionBest AI models for Korean
Korean with correct 존댓말/반말 for KR apps and support.
View collectionBest AI models for Arabic
MSA or dialectal Arabic for MENA products and support.
View collectionBest AI models for Hindi
Hindi (and Hinglish) for CX and products in India.
View collectionAnthropic Claude models on one API
Full Claude family: flagship, sweet spot, fast, and legacy — no separate Anthropic bill required.
View collectionGoogle Gemini (and Veo) models on Geek Hub
Gemini family + Veo: Google chat, vision, image, and video on one gateway.
View collectionGrok (xAI) models via API
xAI stack: Grok chat, reasoning, Build for coding, and Imagine video.
View collectionDeepSeek models: aggressive price and reasoning
DeepSeek Pro and Flash: strong reasoning without a US flagship invoice.
View collectionBest AI models for writing and copy
Prose, essays, newsletters, and scripts with strong-style models.
View collectionBest AI models for math
Math and STEM: reasoning models that fail less on intermediate steps.
View collectionBest AI models for research and synthesis
Research: long context, paper synthesis, and briefs without losing the thread.
View collectionAI models for data analysis
CSV, SQL, EDA, and insight narratives with models that invent fewer numbers.
View collectionAI models for finance and financial analysis
Finance: memos, narrated models, and research with careful tone (not advice).
View collectionAI models for translation (ES ↔ EN and more)
Translation and localization: ES↔EN, glossaries, and brand tone.
View collectionFlux models and pro image generation
Flux Pro/Ultra for photoreal + strong image rivals on the same API.
View collectionVeo 3.1 and Google AI video models
Veo 3.1 as the anchor + video rivals (Sora, Kling, Runway) on one gateway.
View collectionKimi (Moonshot) models for code and 1M context
Kimi K3 (1M) + K2.7 Code: huge context and repos via Moonshot.
View collectionModels with structured outputs / JSON mode
JSON mode and schemas: parseable replies for products and agents.
View collection