Compare vision model APIs for screenshots, documents, product images, and support tickets, with price and capability labels.
Vision APIs are useful for screenshots, receipts, product photos, and ticket attachments. What matters is whether the model accepts the image type you have, whether you need JSON out, and how wrong answers look on your own samples. Price second. The shortlist below is deliberately small — A/B a few vision-capable candidates instead of the whole catalog.
Co so nguon: NextModel capability mapping and OpenRouter input-modality metadata when available. · Cap nhat 2026-07-01
Fit score
Bat dau voi danh sach rut gon, thu prompt thuc te va so sanh chi phi hang thang truoc khi routing production.
Gemini 4 Argon via NextModel: 1M-token context; $2 / 1M tokens in / $10 / 1M tokens out per 1M tokens. Capabilities: streaming, tool, json.
Claude Haiku 5.5 via NextModel: 1M-token context; Starting at $0.1 / 1M tokens in / $0.5 / 1M tokens out per 1M tokens. Capabilities: streaming, tool, json.
Doubao Seedance 2.5 via NextModel: context unpublished; $0.225 / s in / — out per 1M tokens. Capabilities: streaming, vision.
Dreamina-Seedance 2.0 Fast via NextModel: context unpublished; $0.122 / s in / — out per 1M tokens. Capabilities: streaming, vision.
Bang so sanh
Dung giao dien nay de thu hep shortlist production, xay dung chinh sach fallback hoac so sanh kinh te model.
| Model | Provider | Input | Output | Context | Capabilities | Best for | Latency | Status | Source |
|---|---|---|---|---|---|---|---|---|---|
| Gemini 4 Argongoogle/gemini-4-argon | $2 / 1M tokens | $10 / 1M tokens | 1M | StreamingTool callingJSON modeVision | image understanding, multimodal chat | 2974-10589ms | Catalog | Platform curated | |
| Claude Haiku 5.5anthropic/claude-haiku-5.5 | Anthropic | Starting at $0.1 / 1M tokens | $0.5 / 1M tokens | 1M | StreamingTool callingJSON modeVision | image understanding, multimodal chat | 1321-2017ms | Catalog | Platform curated |
| Doubao Seedance 2.5doubao/doubao-seedance-2-5-260628 | Doubao | $0.225 / s | — | — | StreamingVision | image understanding, multimodal chat | 0-0ms | Catalog | Platform curated |
| Dreamina-Seedance 2.0 Fastdreamina/dreamina-seedance-2-0-fast-260128 | Dreamina | $0.122 / s | — | — | StreamingVision | image understanding, multimodal chat | 0-0ms | Catalog | Platform curated |
| Dreamina Seedance 2.5dreamina/dreamina-seedance-2-5-260628 | Dreamina | $0.234 / s | — | — | StreamingVision | image understanding, multimodal chat | 0-0ms | Catalog | Platform curated |
| Gemini 3 Pro Image (Nano Banana Pro) gemini/gemini-3-pro-image | Gemini | $0.314 / image | — | — | StreamingVision | image understanding, multimodal chat | 0-0ms | Catalog | Platform curated |
| Gemini 3.1 Flash Image (Nano Banana 2)gemini/gemini-3.1-flash-image | Gemini | $0.151 / image | — | — | StreamingVision | image understanding, multimodal chat | 0-0ms | Catalog | Platform curated |
| Qwen-Image-2.0 CNqwen/qwen-image-2.0-cn | Qwen | $0.0287 / image | — | — | StreamingVision | image understanding, multimodal chat | 0-0ms | Catalog | Platform curated |
FAQ
Image input support, JSON mode if you need it, latency, output cost, and quality on your real images. Synthetic benchmarks lie more often than people admit.
Sometimes, for light work. Dense documents and high-accuracy extraction usually need a stronger model and a real eval set.
Bang xep hang