← All models
moondream
API MODEL PROFILE / moondream

Moondream: Moondream 3.1

moondream/moondream3.1-9b-a2b
Run model
Catalogue observation · 2026-11-08T21:23:36.187Z · Source: OpenRouter
textText API available

About this API

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...

Provider-supplied description, not an independently validated performance claim.
Input / 1 million tokens$0.3Published prompt tariff
Output / 1 million tokens$2.5Published completion tariff
Context window32,000Published token capacity

These are the catalogue’s token tariffs. Image, audio, video, search and other charges may use different billing units. Zero token rates alone do not mean those services are free. See the complete pricing fields below and the source model page.

Pros and useful characteristics
  • Published input token pricing is at or below the median of catalogue entries with the same output modalities and known tariffs.
  • No separate per-request fee is listed in this catalogue observation. Other modality charges may apply.
  • The upstream model accepts image inputs according to its catalogue. This gateway currently accepts text only.
Cons and practical limits
  • Published output token pricing is above the same-modality catalogue median.
  • The published context window is 32,000 tokens; larger documents may need splitting.
  • Trident Protocol currently limits requests to text content, 24 KB of conversation and at most 4,096 output tokens. Free variants remain subject to upstream rate limits and availability.
  • No independent latency, uptime, reasoning-quality or coding benchmark has been measured by this application.

Assessment uses published fields and unweighted price medians across 437 entries with the same output modalities and known tariffs. It is a cost-and-capability comparison, not a quality ranking or recommendation.

Observed price history

No observations yet. The chart starts with the first recorded data point.

No observations yet. The chart starts with the first recorded data point.

One catalogue observation per hour. History begins when this deployment first observes the model. No backfilled or interpolated prices.

Published model tariffsUSD / 1M tokens
InputOutput
Moondream: Moondream 3.1
$0.30
$2.50

Zero baseline. Published token tariffs, not total multimodal cost or performance scores.

Estimate a request

Published tariff arithmetic; not a guaranteed bill or a gateway reservation.

$0.001550 estimated USD

Excludes unlisted provider-specific charges, taxes and any third-party markup. Context limits, reasoning tokens, caching and provider routing can affect the actual result.

Technical profile
Canonical model
moondream/moondream3.1-9b-a2b-20261013
Separate request tariff
$0
Upstream output limit
4,096
Gateway output limit
4,096
Upstream inputs
text, image
Upstream outputs
text
Knowledge cutoff
Not published

Advertised upstream parameters

include_reasoningmax_completion_tokensmax_tokensreasoningtemperaturetop_p

Upstream capabilities are separate from the text subset exposed by this gateway.

View source model page

Complete published pricing

Source fields are retained exactly, including extra charges and tier overrides. Negative values indicate variable pricing, not rebates. An absent field is not evidence that a service is free.

View OpenRouter pricing fields
{
  "prompt": "0.0000003",
  "completion": "0.0000025"
}
View moondream profile ↗

Use this model with TRIDENT

Connect your wallet, verify your TRIDENT holdings and create an API key. Requests draw from the shared OpenRouter treasury.

Read the API access guide