Define the task
Choose the output first: an answer, an embedding, an image, a transcript or a video. Check the input formats your workflow needs. Reading audio and generating speech are different capabilities.
The SyncDev models directory
Find the right model for the work you need to do. Explore 3,077 profiles, from hosted language models to image, audio and video generators and downloadable checkpoints.
Search model profiles by task, provider and deployment.
Put two models side by side before choosing an API.
Current releases for coding, reasoning and high-volume work. Choose by task, then compare the full cost.
OpenAI
Coding and everyday professional work
Input / 1M tokens
From $2.00
Deployment
Hosted API
Anthropic
Coding and long-running agents
Input / 1M tokens
$4.00
Deployment
Hosted API
Anthropic
Complex reasoning and knowledge work
Input / 1M tokens
$10.00
Deployment
Hosted API
Anthropic
High-volume extraction, classification and routing
Input / 1M tokens
From $0.10
Deployment
Hosted API
Anthropic
Coding and professional workflows
Input / 1M tokens
$2.00
Deployment
Hosted API
OpenAI
Reasoning and demanding workflows
Input / 1M tokens
From $10.00
Deployment
Hosted API
OpenAI
Focused, high-volume tasks
Input / 1M tokens
$0.10
Deployment
Hosted API
xAI
Coding, agents and knowledge work
Input / 1M tokens
$2.00
Deployment
Hosted API
Multimodal tasks and agents
Input / 1M tokens
$0.75
Deployment
Hosted API
Downloadable models to consider for local or managed deployment.
Z.ai / Zhipu AI
Coding and long-horizon tasks
Context
1M
Deployment
Open weights
Moonshot AI
Multimodal reasoning and agents
Context
1.05M
Deployment
Open weights
DeepSeek
Multimodal reasoning and deployment
Context
1.05M
Deployment
Open weights
Mistral AI
Multimodal tasks and open deployment
Context
1M
Deployment
Open weights
Qwen
Large-scale reasoning and coding
Context
262.14K
Deployment
Open weights
Qwen
Multimodal work and efficient inference
Context
262.14K
Deployment
Open weights
Every lab and host with models in the directory.
Choose what you need the model to do. A model can appear in more than one task.
Start with our shortlist, search a model family, or filter by task and deployment. Prices apply to the listed route; a dash means a figure is not recorded.
| GPT-6.1 Sol OpenAI | 75.9 | 1.05M | From $2.00 | From $10.00 | — |
| Claude Opus 5.5 Anthropic | 75.6 | 1M | $4.00 | $20.00 | — |
| Claude Fable 5.1 Anthropic | 75.5 | 1M | $10.00 | $50.00 | — |
| Claude Haiku 5.5 Anthropic | — | 1M | From $0.10 | From $0.50 | — |
| Claude Sonnet 5.5 Anthropic | 75.6 | 1M | $2.00 | $10.00 | — |
| GPT-6 Astra OpenAI | 75.7 | 1.05M | From $10.00 | From $50.00 | — |
| GPT-6 Luna OpenAI | 75.8 | 1.05M | $0.10 | $0.50 | — |
| Grok 4.7 xAI | 58.3 | 500K | $2.00 | $6.00 | — |
| Gemini 3.8 Flash Google | 89.3 | 1.05M | $0.75 | $3.75 | — |
| GLM 5.3 Z.ai / Zhipu AI | 15.0 | 1M | — | — | — |
| Kimi K3 Moonshot AI | 80.8 | 1.05M | $3.00 | $15.00 | — |
| DeepSeek V4.1 Flash DeepSeek | 24.8 | 1.05M | From $0.15 | From $0.60 | — |
| Mistral Large 4 Mistral AI | 66.6 | 1M | From $0.68 | From $2.09 | — |
| Qwen3.8 2.4T A95B Qwen | 35.2 | 262.14K | — | — | — |
| Qwen3.8 Flash Next Qwen | 45.2 | 262.14K | — | — | — |
| GPT-6 Sol OpenAI | 75.8 | 1.05M | $2.00 | $10.00 | — |
| Qwen3.8 Max Alibaba | 75.7 | 1M | $2.00 | $6.00 | — |
| Muse Spark 1.3 Meta | — | 1.05M | From $1.25 | From $4.25 | — |
| GLM 5.3 Flash Z.ai / Zhipu AI | 61.9 | 1M | — | — | — |
| Qwen3.8 27B Qwen | 45.2 | 262.14K | — | — | — |
| Llama 4 Maverick 17B Instruct Meta | 48.5 | 1M | $0.27 | $0.85 | — |
| Gemini 4 Argon Google | — | 1.05M | From $2.00 | From $10.00 | — |
| Muse Wonder Meta | — | 1.05M | From $4.00 | From $20.00 | — |
| Muse Wonder Full Meta | — | 1.05M | — | — | — |
| Qwen-Image-2.1-Turbo Alibaba | — | — | — | — | — |
| Phocinae Largha 150M v1 phocinae | — | 8.19K | — | — | — |
| Drex 2.0 Nace AI | — | 8.19K | From $0.06 | From $0.00 | — |
| Gemini Nano Banana 2.1 Google | 62.9 | 131.07K | — | — | — |
| Qwen-Image-2.1-Pro Alibaba | — | — | — | — | — |
| Grok Imagine Video 1.5 Lite xAI | 10.0 | — | — | — | — |
| MAI-Transcribe-2-Streaming Microsoft | — | — | — | — | — |
| MAI-Voice-2.1 Microsoft | — | — | — | — | — |
| MAI-Voice-2.1-Flash Microsoft | — | — | — | — | — |
| Ling 3.1 Flash inclusionAI | 51.3 | 262.14K | — | — | — |
| Liquid d1 Liquid AI | 10.0 | — | — | — | — |
| Drex 1.5 Nace AI | — | 131.07K | From $0.05 | From $0.00 | — |
| MiniMax M3.1 Flash Preview MiniMax | 61.7 | 1M | — | — | — |
| LongCat-2.5-Preview Meituan | — | 1M | — | — | — |
| Qwen 3.8 Max Prime Alibaba | — | 1M | — | — | — |
| Gemini 3.8 Flash TTS Google | 17.9 | 8.19K | — | — | — |
| Gemini 3.8 Flash-Lite TTS Google | 17.9 | 8.19K | — | — | — |
| MiMo V2.6 Flash RL Xiaomi | 66.9 | 1.05M | From $0.14 | From $0.28 | — |
| MiMo V2.6 Pro RL Xiaomi | 66.9 | 1.05M | From $0.435 | From $0.87 | — |
| MiMo-V2.6-Distill-Qwen-9B Xiaomi | — | — | — | — | — |
| Viv Fast Vivgrid | — | 1M | $0.13 | $0.40 | — |
| Qwen-Audio-3.1-Realtime-Plus Alibaba | — | 262.14K | — | — | — |
| Grok Voice Transcribe 2.0 xAI | — | — | — | — | — |
| Pareto Unbiased AI | — | 262.14K | — | — | — |
| Qwen3.8 LiveTranslate Flash Realtime Alibaba | — | — | — | — | — |
| Qwen3.8 Omni Flash Qwen | 51.7 | 1M | — | — | — |
| Arrow 2 QuiverAI | — | 131.07K | — | — | — |
| Arrow 2 Telos QuiverAI | — | 1.05M | — | — | — |
| Step 5 Preview StepFun | 74.2 | 1M | $1.00 | $2.70 | — |
| Gemini 3.8 Live Google | 74.3 | 131.07K | — | — | — |
| Jev TypeSafe | — | 64K | From $0.042 | From $0.00 | — |
| Qwen Image 2.1 Qwen | 10.0 | — | — | — | — |
| Schematron V2 Small inference-net | — | 128K | — | — | — |
| Schematron V2 Turbo inference-net | — | 128K | — | — | — |
| Fugu Max Sakana AI | — | 1M | — | — | — |
| Fugu Ultra v2 Sakana AI | — | 1M | — | — | — |
1–60 of 3077 models · select a column to sort
About this order: Recommended shows our editorial shortlist first, followed by recent releases. The spec score is a 0–100 figure SyncDev calculates from each vendor's published specifications — context window, reasoning support, input modalities, tool calling, maximum output and how recently the model shipped. It measures breadth of capability, not benchmark performance, so a higher-scoring model is not automatically the better choice for your task.
Choose the output first: an answer, an embedding, an image, a transcript or a video. Check the input formats your workflow needs. Reading audio and generating speech are different capabilities.
For language models, include input, generated output, reasoning and retries. For media, check the billing unit, resolution and duration. Open weights need a runtime and hardware even when the download is free.
Give each candidate the same real examples and acceptance criteria. Measure accuracy, reliability and the time to a usable result. A larger context window or a broader specification sheet alone cannot tell you which model will work best.