Side by side
See the difference.
Scores, capabilities and costs for the models you’re considering.
Scroll to compare models →
| 3 models | bytedance-seedByteDance Seed: Seedream 4.5Remove × | ByteDanceSeed-2.0-codeRemove × | moondreamMoondream: Moondream 3 PreviewRemove × |
|---|---|---|---|
| What’s different | Seedream 4.5 is the latest in-house image generation model developed by ByteDance. Compared with Seedream 4.0, it delivers comprehensive improvements, especially in editing consistency, including better… | A coding model optimized for real-world development environments, with reliable tool use in common IDEs such as Claude Code. It delivers strong front-end performance and supports Skills. | 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… |
| Date | Dec 5, 2025Released | Apr 29, 2026Listed | Oct 15, 2026Listed |
API prices USD · provider configuration shown below | |||
| Listed rate | $0.04/ imagefal · all routes ↗ | $0.50 in / $3.00 out/ 1M tokensDeepInfra · all routes ↗ | $0.30 in / $2.50 out/ 1M tokensPrimeIntellect · via OpenRouter · all routes ↗primeintellect/fp8 · quantization=fp8 |
| Provider endpoints | 2 endpoints → | 1 endpoint → | 2 endpoints → |
Benchmarks Each result retains its evaluation configuration | |||
| AA Image ↗2.1 | 1,026.35 ↗Text-to-image · all categoriesIndependent result95% CI 1,014.35–1,038.35 · n=6,135 | Not reported | Not reported |
Specifications | |||
| Context window | 4.1K | 256K | 32K |
| Max output | 3.7K | — | 4.1K |
| Parameters | — | — | — |
| Image input | Supported | Not reported | Supported |
| Reasoning | Not reported | Not reported | Supported |
| Open weights | Not reported | Not reported | Supported |