Beta
Use case · Vision

Cheapest vision LLMs

The cheapest multimodal models that accept images. Ranked by input token price, with notes on per-image billing and OCR quality.

Models30
Cheapest$0.00
TypeMultimodal
What this page is
This page lists every multimodal model (vision capable) with priced API access, ranked cheapest first. Note that per-image billing can diverge from per-token billing: a single high-res image may cost hundreds of tokens depending on the provider. Use the input price below as a baseline, then check the model detail page for image-specific pricing notes. Ideal for OCR, document processing, chart reading, and visual QA.

Multimodal models only, cheapest first.

#ModelIn $/1MOut $/1MType
1openrouter logoFree Models Router$0.00$0.00Closed
2Google DeepMind logoGemma 3 12B (free)$0.00$0.00OSS
3Google DeepMind logoGemma 3 27B (free)$0.00$0.00OSS
4Google DeepMind logoGemma 3 4B (free)$0.00$0.00OSS
5Google DeepMind logoGemma 4 26B A4B (free)$0.00$0.00OSS
6Google DeepMind logoGemma 4 31B (free)$0.00$0.00OSS
7Mistral AI logoMistral Small 3.1 24B (free)$0.00$0.00OSS
8NVIDIA logoNemotron Nano 12B 2 VL (free)$0.00$0.00OSS
9Alibaba Qwen logoQwen3.6 Plus (free)$0.00$0.00Closed
10Google DeepMind logoGemma 3 12B$0.04$0.13OSS
11Google DeepMind logoGemma 3 4B$0.04$0.08OSS
12OpenAI logoGPT-5 Nano$0.05$0.40Closed
13Alibaba Qwen logoQwen3.5-9B$0.05$0.15OSS
14Amazon logoNova Lite 1.0$0.06$0.24Closed
15Alibaba Qwen logoQwen3.5-Flash$0.07$0.26OSS
16Google DeepMind logoGemini 2.0 Flash Lite$0.07$0.30Closed
17Mistral AI logoMistral Small 3.2 24B$0.07$0.20OSS
18ByteDance logoSeed 1.6 Flash$0.07$0.30Closed
19Google DeepMind logoGemma 3 27B$0.08$0.16OSS
20Google DeepMind logoGemma 4 26B A4B $0.08$0.35OSS
21Meta logoLlama 4 Scout$0.08$0.30OSS
22Alibaba Qwen logoQwen3 VL 8B Instruct$0.08$0.50OSS
23Google DeepMind logoGemini 2.0 Flash$0.10$0.40Closed
24Google DeepMind logoGemini 2.5 Flash Lite$0.10$0.40Closed
25Google DeepMind logoGemini 2.5 Flash Lite Preview 09-2025$0.10$0.40Closed
26OpenAI logoGPT-4.1 Nano$0.10$0.40Closed
27Mistral AI logoMinistral 3 3B 2512$0.10$0.10OSS
28rekaai logoReka Edge$0.10$0.10OSS
29ByteDance logoSeed-2.0-Mini$0.10$0.40Closed
30bytedance logoUI-TARS 7B $0.10$0.20OSS
Cheapest
Free Models Router
$0.00/M
$ per 1M input tokens
Why the gap

Premium vision models pay for higher resolution encoders, better chart and table reading, and longer context for multi-page documents. For batch OCR, the cheap end is often good enough.

Most expensive
Gemini 2.0 Flash
$0.10/M
$ per 1M input tokens
Most providers bill images as a fixed token count based on resolution. OpenAI bills ~255 tokens per low-res image, 765+ for high-res. Anthropic bills by image dimensions. Google bills flat per image. Always check the provider docs before budgeting.