All models
Moondream
Moondream 3 Preview balancedruns locally
moondream/moondream3-preview
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...
Input / 1M
$0.30
USD per million tokens
Output / 1M
$2.50
USD per million tokens
Est. cost / token
$0.00000085
$0.85 per 1M, 3:1 mix
Context window
32K
max output 4K
Major features
VisionOpen weightsReasoning
Task suitability (heuristic 1–5, not a benchmark)
- coding4
- reasoning4
- math4
- summarization4
- extraction4
- classification4
- chat4
- creative4
- translation4
- vision4
- long context2
- agentic2
Run it locally
Will it run on my machine? →| Quantization | Quality | Memory needed (8K ctx) |
|---|---|---|
| Q2_K | poor | 5.8 GB |
| Q3_K_M | fair | 6.5 GB |
| Q4_K_Mrecommended | good | 7.7 GB |
| Q5_K_M | very good | 8.7 GB |
| Q6_K | excellent | 9.8 GB |
| Q8_0 | near-lossless | 12.1 GB |
| F16 | lossless | 21.3 GB |
Estimates.
Price history
| Recorded | Input | Output | Cached | Source |
|---|---|---|---|---|
| 2026-11-05 03:20 | $0.30 | $2.50 | — | openrouter |
A row is added whenever a sync detects a price change. Vendor pricing source
Pricing (USD per 1M tokens)
- Input
- $0.30
- Output
- $2.50
- Blended (3:1)
- $0.85
Details
- Company
- Moondream
- Version
- 3
- Family
- Moondream
- Released
- 2026-10-15
- Parameters
- 9.3B
- Max output
- 4K
- Open weights
- Yes
- Inputs
- image, text
- Last synced
- 21 min ago