GPT-5.4 Mini — latest gen fast and affordable

Newest compact reasoning model with tool use support

GPT-5.1 — improved reasoning and longer context window

GPT-5 flagship model with advanced reasoning and longer context

GPT-5.4 Nano — ultra-fast, low-latency for classification, extraction, and sub-agents

GPT-5.3 Chat — previous generation flagship chat model

GPT-5.1 chat-optimized variant

GPT-5.1 Codex — advanced code generation and analysis

GPT-5.1 Codex Mini — fast and affordable code generation

GPT-5 chat-optimized variant

GPT-5 Mini — affordable with strong reasoning

GPT-5 Nano — fastest and cheapest in the GPT-5 family

GPT-5 Pro — maximum capability, best for complex tasks

GPT-5 Codex — advanced code generation and analysis

Latest reasoning model with improved speed and accuracy

Flagship model with 200K context, best for complex reasoning and coding

Fast and affordable, great balance of speed and intelligence

Fastest and cheapest, ideal for simple tasks and classification

Previous flagship with vision, strong all-around performance

Compact model optimized for speed and cost efficiency

Multimodal model supporting audio input and output

GPT-4o Mini TTS — expressive, natural speech with emotion and tone control

GPT-4o powered transcription — more accurate than Whisper, supports structured output

Fast and affordable GPT-4o Mini transcription

GPT-4o transcription with speaker diarization — identifies who said what

Compact embedding model, 1536 dimensions

Large embedding model, up to 3072 dimensions

Legacy embedding model, 1536 dimensions

Standard text-to-speech, fast and natural sounding

High-definition text-to-speech with premium voice quality

Industry-leading speech-to-text transcription

GPT-OSS 120B — open-source large language model

GPT-OSS 20B — compact open-source model

Latest Opus with improved coding and reduced cost

Latest Sonnet, top choice for Claude Code and Cursor

Fast and capable, great for real-time applications

Sonnet 4 with extended thinking for deeper reasoning

Claude Opus 4.5 via Bedrock, top-tier reasoning and creativity

Claude Sonnet 4.5 via Bedrock, strong balanced performance

Most capable Claude yet (2026-04-16): 1M context, 128K output, agentic workflows

Claude Opus 4.1 via Bedrock, premium reasoning model

Balanced Claude with strong coding and reasoning abilities

Previous Haiku generation, compact and efficient

Google most capable model with 1M context window

Fast Gemini with strong reasoning at low cost

Gemini 3 Pro Preview — next-gen reasoning

Gemini 3 Pro Image — highest quality image generation & editing, up to 4K

Gemini 3 Flash Preview — fast next-gen model

Gemini 2.5 Flash Lite — lightweight and fast

Gemini 2.5 Flash Image — vision and image generation

Gemini 3.1 Pro Preview — latest capabilities

Gemini 3.1 Flash Image — vision and image generation

Gemini 3.1 Flash Lite — ultra-low cost for high-throughput tasks

Gemini embedding model for text retrieval

Latest DeepSeek V3.2 with hybrid thinking mode

DeepSeek R1 snapshot from May 2025

Chain-of-thought reasoning model rivaling o1

DeepSeek V3.1 with hybrid thinking mode

DeepSeek V3 text generation model

DeepSeek R1 reasoning-only model

Open-source powerhouse, strong coding and math skills

Qwen 3.6 Plus — latest flagship, rivals Claude Opus 4.5 on benchmarks

Specialized coding model by Qwen

Next-gen Qwen coding model, top tier for code tasks

Qwen3 Coder 30B, compact coding specialist via Bedrock

Qwen3 Coder Flash — fastest, cheapest coding model

Qwen3 Coder 480B — flagship coding model, largest in the series

Qwen reasoning model, rivals DeepSeek-R1

Qwen 3 flagship, top reasoning and coding

Qwen 3.5 Plus with 1M context

Ultra-fast Qwen 3.5 with 1M context

Qwen 3.5 Omni native multimodal — text, image, audio, video

Qwen 3.5 397B MoE flagship, top-tier reasoning

Qwen vision-language model

Qwen3 Next 80B via Bedrock, efficient MoE architecture

Qwen3 VL 235B, large vision-language model via Bedrock

Qwen3 32B dense model via Bedrock, strong all-rounder

Qwen3 TTS Flash — fast text-to-speech with natural voice

Qwen3 TTS Instruct — instruction-controlled speech synthesis

Qwen 3.5 Omni Flash — fast multimodal (text+image+audio)

Qwen 3.5 Omni Plus — premium multimodal understanding

Qwen3 VL Flash — fast and cheap visual understanding

Balanced Qwen model via direct API

Always-latest Qwen Plus snapshot

Fast Qwen model, deprecated in favor of Qwen Flash

10M ultra-long context for massive documents

Qwen flagship via direct API

Specialized coding model with 32B parameters

Qwen vision-language model via direct API

Ultra-fast Qwen Flash, upgraded to Qwen3.5

Always-latest Qwen Max snapshot

Qwen VL Plus for vision-language tasks

Wan 2.7 Image — affordable image generation

Wan 2.7 Image Pro — high-quality image generation

CosyVoice v2 text-to-speech via Bailian

SenseVoice speech-to-text via Bailian

Paraformer v2 speech recognition via Bailian

Wan 2.7 Text-to-Video

Wan 2.6 Text-to-Video

Wan 2.7 Image-to-Video

Wan 2.6 Image-to-Video

Wan 2.6 Image-to-Video Flash — fast generation

QVQ-Max — flagship visual reasoning model, deep visual understanding

QVQ-Plus — balanced visual reasoning, cost-effective

Qwen VL OCR — specialized document OCR, table/form extraction

Qwen Image 2.0 — text-to-image generation

Qwen Image 2.0 Pro — high-quality text-to-image

Qwen Image Max — best quality text-to-image

Qwen Image Edit Plus — conversational image editing, Chinese & English prompts

Qwen Image Edit Max — premium conversational image editing

Qwen Math Plus — specialized mathematical reasoning

Qwen Math Turbo — fast math reasoning at lower cost

Qwen Omni Turbo — text, image, and audio understanding

Z-Image Turbo — lightweight fast image generation

Qwen MT Plus — professional machine translation

Qwen MT Turbo — fast translation at lower cost

GLM-5.1 latest — improved coding and reasoning, 94% of Claude Opus 4.6 coding