How do I use Moondream models through the OpenRouter API?

OpenRouter serves 3 Moondream models behind one OpenAI-compatible API. Create an OpenRouter API key, point your client at https://openrouter.ai/api/v1, and set the model to an ID such as moondream/parakeet-ultra. The quickstart has request examples for every supported SDK.

Which Moondream model has the largest context window?

Moondream 3 Preview has the largest context window of any Moondream model on OpenRouter, accepting up to 32,000 tokens per request.

What is the newest Moondream model on OpenRouter?

Parakeet Ultra is the most recently added Moondream model on OpenRouter, listed on November 3, 2026.

Moondream

Access 3 Moondream models through the OpenRouter unified API including Parakeet Ultra, Moondream 3 Preview, and Moondream 3.1. Compare pricing, context windows, benchmarks, and capabilities between different Moondream models.

Moondream tokens processed on OpenRouter

Moondream: Parakeet UltraParakeet Ultra
1.28M characters
Parakeet Ultra is Moondream's post-trained version of NVIDIA's Parakeet TDT 0.6B v3 speech-to-text model, with the same architecture, tokenizer, and 0.6B parameters. It lowers word error rate on the Open ASR Leaderboard English sets (5.80% vs 6.26%) and the 25-language FLEURS set (9.55% vs 11.62%), with larger gains on noisy and long-form audio. Supports 25 European languages and returns transcribed text with segment and word timestamps.
by moondreamNov 3, 2026$0.000017/second
Moondream: Moondream 3 PreviewMoondream 3 Preview
61.9M tokens
Moondream 3 Preview is a vision-language model from Moondream with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with a 32K-token context and an optional reasoning mode.
by moondreamOct 15, 202632K context$0.30/M input tokens$2.50/M output tokens
Moondream: Moondream 3.1Moondream 3.1
214M tokens
Moondream 3.1 is a vision-language model from Moondream, the successor to Moondream 3 Preview, with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with a 32K-token context and an optional reasoning mode.
by moondreamOct 15, 202632K context$0.30/M input tokens$2.50/M output tokens

Moondream

Access 3 Moondream models through the OpenRouter unified API including Parakeet Ultra, Moondream 3 Preview, and Moondream 3.1. Compare pricing, context windows, benchmarks, and capabilities between different Moondream models.

Moondream tokens processed on OpenRouter

Moondream: Parakeet UltraParakeet Ultra
1.28M characters
Parakeet Ultra is Moondream's post-trained version of NVIDIA's Parakeet TDT 0.6B v3 speech-to-text model, with the same architecture, tokenizer, and 0.6B parameters. It lowers word error rate on the Open ASR Leaderboard English sets (5.80% vs 6.26%) and the 25-language FLEURS set (9.55% vs 11.62%), with larger gains on noisy and long-form audio. Supports 25 European languages and returns transcribed text with segment and word timestamps.
by moondreamNov 3, 2026$0.000017/second
Moondream: Moondream 3 PreviewMoondream 3 Preview
61.9M tokens
Moondream 3 Preview is a vision-language model from Moondream with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with a 32K-token context and an optional reasoning mode.
by moondreamOct 15, 202632K context$0.30/M input tokens$2.50/M output tokens
Moondream: Moondream 3.1Moondream 3.1
214M tokens
Moondream 3.1 is a vision-language model from Moondream, the successor to Moondream 3 Preview, with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and visual reasoning over images, with a 32K-token context and an optional reasoning mode.
by moondreamOct 15, 202632K context$0.30/M input tokens$2.50/M output tokens

Moondream

Documentation

Frequently asked questions

Moondream

Documentation

Frequently asked questions