Compare AI model APIs by modality, provider, price, and context window
Browse 5,186 model API records across 141 providers. Each section below groups models by input and output modality, so you can compare text, image, audio, video, PDF, and multimodal APIs in one place.
Model vendors and gateway providers
Normalized model API records
Records with input or output costs
Input -> output groupings
Filter by input and output modality
Choose an input type and an output type to quickly narrow the directory. Add useful rows to the browser compare list, then open Compare to start with those models selected.
AI model API comparison tables by modality
Tables are sorted by context window within each modality group, then by provider and model name. Prices are shown as normalized numeric values, displayed as per 1M units where applicable.
2,508 model API records in this modality group.
| Compare | Provider | Model | Family | Input | Output | Context | Max output | Input price | Output price | Reasoning | Tools | Updated |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Alibaba (China) alibaba-cn | Qwen Long qwen-long | qwen | text | text | 10,000,000 | 8,192 | $0.072 / 1M | $0.287 / 1M | No | Yes | Jan 25, 2025 | |
NanoGPT nano-gpt | Gemini 1.5 Flash google/gemini-flash-1.5 | gemini-flash | text | text | 2,000,000 | 8,192 | $0.0748 / 1M | $0.306 / 1M | No | No | May 14, 2024 | |
OpenRouter openrouter | Pareto Code Router openrouter/pareto-code | Unknown | text | text | 2,000,000 | 200,000 | Unknown | Unknown | No | No | Apr 21, 2026 | |
Qiniu qiniu-ai | x-AI/Grok-4.1-Fast x-ai/grok-4.1-fast | Unknown | text | text | 2,000,000 | 2,000,000 | Unknown | Unknown | Yes | Yes | Nov 20, 2025 | |
Requesty requesty | Grok 4 Fast xai/grok-4-fast | grok | text | text | 2,000,000 | 64,000 | $0.2 / 1M | $0.5 / 1M | Yes | Yes | Sep 19, 2025 | |
SiliconFlow (China) siliconflow-cn | deepseek-ai/DeepSeek-V4-Pro deepseek-ai/DeepSeek-V4-Pro | deepseek-thinking | text | text | 1,049,000 | 393,000 | $1.74 / 1M | $3.48 / 1M | Yes | Yes | Apr 24, 2026 | |
Kilo Gateway kilo | Owl Alpha openrouter/owl-alpha | Unknown | text | text | 1,048,756 | 262,144 | $0 / 1M | $0 / 1M | Yes | Yes | Apr 30, 2026 | |
NanoGPT nano-gpt | OWL owl | Unknown | text | text | 1,048,756 | 262,144 | $0.1 / 1M | $0.3 / 1M | No | Yes | May 1, 2026 | |
OpenRouter openrouter | Owl Alpha openrouter/owl-alpha | alpha | text | text | 1,048,756 | 262,144 | $0 / 1M | $0 / 1M | No | Yes | Apr 28, 2026 | |
AIHubMix aihubmix | Coding Xiaomi MiMo-V2.5-Pro coding-xiaomi-mimo-v2.5-pro | mimo-v2.5-pro | text | text | 1,048,576 | 131,072 | $0.2 / 1M | $0.6 / 1M | Yes | Yes | May 13, 2026 | |
AIHubMix aihubmix | Xiaomi MiMo-V2.5-Pro xiaomi-mimo-v2.5-pro | mimo-v2.5-pro | text | text | 1,048,576 | 131,072 | $1.1 / 1M | $3.3 / 1M | Yes | Yes | May 13, 2026 | |
AIHubMix aihubmix | Xiaomi MiMo-V2.5-Pro (free) xiaomi-mimo-v2.5-pro-free | mimo-v2.5-pro | text | text | 1,048,576 | 131,072 | $0 / 1M | $0 / 1M | Yes | Yes | May 13, 2026 | |
Alibaba (China) alibaba-cn | Qwen3 Coder Plus qwen3-coder-plus | qwen | text | text | 1,048,576 | 65,536 | $1 / 1M | $5 / 1M | No | Yes | Jul 23, 2025 | |
Alibaba alibaba | Qwen3 Coder Plus qwen3-coder-plus | qwen | text | text | 1,048,576 | 65,536 | $1 / 1M | $5 / 1M | No | Yes | Jul 23, 2025 | |
Cortecs cortecs | DeepSeek V4 Flash deepseek-v4-flash | deepseek-flash | text | text | 1,048,576 | 384,000 | $0.133 / 1M | $0.266 / 1M | Yes | Yes | Apr 24, 2026 | |
Cortecs cortecs | DeepSeek V4 Pro deepseek-v4-pro | deepseek-thinking | text | text | 1,048,576 | 384,000 | $1.553 / 1M | $3.106 / 1M | Yes | Yes | Apr 24, 2026 | |
CrofAI crof | MiMo-V2.5-Pro mimo-v2.5-pro | mimo | text | text | 1,048,576 | 131,072 | $0.4 / 1M | $0.8 / 1M | Yes | Yes | Apr 22, 2026 | |
Deep Infra deepinfra | DeepSeek V4 Flash deepseek-ai/DeepSeek-V4-Flash | deepseek-flash | text | text | 1,048,576 | 16,384 | $0.1 / 1M | $0.2 / 1M | Yes | Yes | Apr 24, 2026 | |
Deep Infra deepinfra | DeepSeek V4 Pro deepseek-ai/DeepSeek-V4-Pro | deepseek-thinking | text | text | 1,048,576 | 16,384 | $1.3 / 1M | $2.6 / 1M | Yes | Yes | Apr 24, 2026 | |
Deep Infra deepinfra | MiMo-V2.5-Pro XiaomiMiMo/MiMo-V2.5-Pro | mimo | text | text | 1,048,576 | 16,384 | $1 / 1M | $3 / 1M | Yes | Yes | Apr 22, 2026 | |
DigitalOcean digitalocean | DeepSeek V4 Pro deepseek-v4-pro | deepseek-thinking | text | text | 1,048,576 | 393,216 | $1.74 / 1M | $3.48 / 1M | Yes | Yes | Apr 24, 2026 | |
GMI Cloud gmicloud | DeepSeek V4 Pro deepseek-ai/DeepSeek-V4-Pro | deepseek-thinking | text | text | 1,048,576 | 384,000 | $1.392 / 1M | $2.784 / 1M | Yes | Yes | Apr 24, 2026 | |
Hugging Face huggingface | DeepSeek V4 Pro deepseek-ai/DeepSeek-V4-Pro | deepseek-thinking | text | text | 1,048,576 | 393,216 | $0.435 / 1M | $0.87 / 1M | Yes | Yes | Apr 24, 2026 | |
Kilo Gateway kilo | DeepSeek: DeepSeek V4 Flash deepseek/deepseek-v4-flash | Unknown | text | text | 1,048,576 | 384,000 | $0.14 / 1M | $0.28 / 1M | Yes | Yes | May 1, 2026 | |
Kilo Gateway kilo | DeepSeek: DeepSeek V4 Pro deepseek/deepseek-v4-pro | Unknown | text | text | 1,048,576 | 384,000 | $0.435 / 1M | $0.87 / 1M | Yes | Yes | May 1, 2026 | |
Kilo Gateway kilo | Xiaomi: MiMo V2.5 Pro xiaomi/mimo-v2.5-pro | mimo | text | text | 1,048,576 | 131,072 | $1 / 1M | $3 / 1M | Yes | Yes | Apr 22, 2026 | |
Kilo Gateway kilo | Xiaomi: MiMo-V2-Pro xiaomi/mimo-v2-pro | mimo | text | text | 1,048,576 | 131,072 | $1 / 1M | $3 / 1M | Yes | Yes | Mar 18, 2026 | |
LLM Gateway llmgateway | MiMo-V2-Pro mimo-v2-pro | mimo | text | text | 1,048,576 | 131,072 | $1 / 1M | $3 / 1M | Yes | Yes | Mar 18, 2026 | |
LLM Gateway llmgateway | MiMo-V2.5-Pro mimo-v2.5-pro | mimo | text | text | 1,048,576 | 131,072 | $1 / 1M | $3 / 1M | Yes | Yes | Apr 22, 2026 | |
LLM Gateway llmgateway | Qwen3 Coder Plus qwen3-coder-plus | qwen | text | text | 1,048,576 | 65,536 | $1 / 1M | $5 / 1M | No | Yes | Jul 23, 2025 | |
NanoGPT nano-gpt | DeepSeek Latest deepseek/deepseek-latest | Unknown | text | text | 1,048,576 | 384,000 | $1.1 / 1M | $2.2 / 1M | Yes | Yes | May 3, 2026 | |
NanoGPT nano-gpt | DeepSeek V4 Flash deepseek/deepseek-v4-flash | Unknown | text | text | 1,048,576 | 384,000 | $0.14 / 1M | $0.28 / 1M | Yes | Yes | Apr 24, 2026 | |
NanoGPT nano-gpt | DeepSeek V4 Flash (Thinking) deepseek/deepseek-v4-flash:thinking | Unknown | text | text | 1,048,576 | 384,000 | $0.14 / 1M | $0.28 / 1M | Yes | Yes | Apr 24, 2026 | |
NanoGPT nano-gpt | DeepSeek V4 Pro deepseek/deepseek-v4-pro | Unknown | text | text | 1,048,576 | 384,000 | $1.1 / 1M | $2.2 / 1M | Yes | Yes | Apr 24, 2026 | |
NanoGPT nano-gpt | DeepSeek V4 Pro (Thinking) deepseek/deepseek-v4-pro:thinking | Unknown | text | text | 1,048,576 | 384,000 | $1.1 / 1M | $2.2 / 1M | Yes | Yes | Apr 24, 2026 | |
NanoGPT nano-gpt | DeepSeek V4 Pro Cheaper deepseek/deepseek-v4-pro-cheaper | Unknown | text | text | 1,048,576 | 384,000 | $0.435 / 1M | $0.87 / 1M | Yes | Yes | Apr 25, 2026 | |
NanoGPT nano-gpt | DeepSeek V4 Pro Cheaper (Thinking) deepseek/deepseek-v4-pro-cheaper:thinking | Unknown | text | text | 1,048,576 | 384,000 | $0.435 / 1M | $0.87 / 1M | Yes | Yes | Apr 25, 2026 | |
NanoGPT nano-gpt | MiMo V2 Pro xiaomi/mimo-v2-pro | Unknown | text | text | 1,048,576 | 131,072 | $1 / 1M | $3 / 1M | Yes | Yes | Mar 19, 2026 | |
NanoGPT nano-gpt | MiMo V2.5 Pro xiaomi/mimo-v2.5-pro | Unknown | text | text | 1,048,576 | 131,072 | $0.435 / 1M | $0.87 / 1M | Yes | Yes | Apr 22, 2026 | |
NovitaAI novita-ai | DeepSeek V4 Flash deepseek/deepseek-v4-flash | deepseek-flash | text | text | 1,048,576 | 393,216 | $0.14 / 1M | $0.28 / 1M | Yes | Yes | Apr 24, 2026 | |
NovitaAI novita-ai | DeepSeek V4 Pro deepseek/deepseek-v4-pro | deepseek-thinking | text | text | 1,048,576 | 393,216 | $1.69 / 1M | $3.38 / 1M | Yes | Yes | Apr 24, 2026 | |
NovitaAI novita-ai | MiMo-V2-Pro xiaomimimo/mimo-v2-pro | mimo | text | text | 1,048,576 | 131,072 | $2 / 1M | $6 / 1M | Yes | Yes | May 27, 2026 | |
NovitaAI novita-ai | MiMo-V2.5-Pro xiaomimimo/mimo-v2.5-pro | mimo | text | text | 1,048,576 | 131,072 | $2 / 1M | $6 / 1M | Yes | Yes | May 27, 2026 | |
Nvidia nvidia | DeepSeek V4 Flash deepseek-ai/deepseek-v4-flash | deepseek-flash | text | text | 1,048,576 | 393,216 | $0.14 / 1M | $0.28 / 1M | Yes | Yes | Apr 24, 2026 | |
Nvidia nvidia | DeepSeek V4 Pro deepseek-ai/deepseek-v4-pro | deepseek-thinking | text | text | 1,048,576 | 393,216 | $0.435 / 1M | $0.87 / 1M | Yes | Yes | Apr 24, 2026 | |
Ollama Cloud ollama-cloud | deepseek-v4-flash deepseek-v4-flash | deepseek-flash | text | text | 1,048,576 | 1,048,576 | Unknown | Unknown | Yes | Yes | Apr 24, 2026 | |
Ollama Cloud ollama-cloud | deepseek-v4-pro deepseek-v4-pro | deepseek-thinking | text | text | 1,048,576 | 1,048,576 | Unknown | Unknown | Yes | Yes | Apr 24, 2026 | |
Ollama Cloud ollama-cloud | nemotron-3-nano:30b nemotron-3-nano:30b | nemotron | text | text | 1,048,576 | 131,072 | Unknown | Unknown | Yes | Yes | Jan 19, 2026 | |
OpenCode Go opencode-go | MiMo V2 Pro mimo-v2-pro | mimo-v2-pro | text | text | 1,048,576 | 128,000 | $1 / 1M | $3 / 1M | Yes | Yes | Mar 18, 2026 | |
OpenCode Go opencode-go | MiMo V2.5 Pro mimo-v2.5-pro | mimo-v2.5-pro | text | text | 1,048,576 | 128,000 | $1.74 / 1M | $3.48 / 1M | Yes | Yes | Apr 22, 2026 |
AI Model API Directory FAQ
Answers to common questions about filtering model APIs, interpreting pricing, and saving models for comparison.
How do I find an AI model API for a specific input and output type?+
Use the input and output modality filters to find models that match the workflow you need. For example, choose text input and image output to narrow the directory to text-to-image models, or choose image input and text output for vision-language models. The directory updates the visible groups immediately, so you can scan only the models that support the required combination. After filtering, add promising models to the compare list for a side-by-side check.
Which cost fields should I check before choosing a model API?+
Check both input cost and output cost before choosing a model API. A model with cheap input can still become expensive if your use case generates long answers, images, audio, or other high-cost outputs. Treat the directory pricing as screening data, then verify the provider billing unit, discounts, and current terms before production use. For a clearer decision, add at least two models to the comparison page and look for the lowest highlighted cost cells.
Can I compare models from this directory immediately?+
Yes, use the Add button in any model row to save that model to your browser compare list. The compare list stays in your browser, so the Compare page can load those models automatically on the same device and browser. Add two or more models when you want a meaningful comparison of cost, limits, modalities, and capability flags. Remove models from the comparison page when they are no longer relevant.
Should I pick the cheapest AI model API in the directory?+
Do not pick a model only because one cost field is the cheapest. Start with modality fit, then check context window, output limit, input cost, output cost, reasoning support, tool calling, and open-weights status. The cheapest model can be the wrong choice if it cannot accept your inputs, produce the required output, or handle the prompt length. Use cost as a shortlist signal, not as the final model selection rule.
How fresh is the model API information?+
The page shows an update date near the top and model rows include provider update fields when available. Use those dates to judge whether a model record is recent enough for early screening. Model availability, names, prices, and limits can change quickly, so confirm critical details with the provider before launch. This directory is best used for discovery and comparison, not as a contract source.
Does this directory rank AI model quality?+
No, this directory does not rank model quality or reliability. It compares technical availability fields such as modality, context window, output limit, pricing, reasoning support, tool calling, and open-weights status. Benchmark scores, production reliability, license terms, and real workload quality need separate evaluation. Use this page to find candidates, then test shortlisted models against your own prompts and constraints.
Limits
Context and output limits help screen models for long documents, coding agents, and large retrieval workflows.
Pricing
Input and output costs are useful for budget screening, but provider-specific pricing should be verified before production use.
Capabilities
Reasoning, tool calling, attachment, and open-weights flags are technical availability signals, not quality scores.
- Commercial license terms are not included in the source payload.
- Run counts, popularity, benchmark scores, and reliability rankings are not included.
- Gateway pricing can differ from original vendor pricing.
- This directory compares technical availability, not model quality.