Call Sheet · Image and pre-production

The best AI models for text in images

Titles and words that have to read, inside the picture. These are Call Sheet's picks for this job, in order, with the reason for each. Call Sheet is a free tool by Melissa Beery that ranks 101 image, video and sound models for a creative job. Its data was last updated on 20 October 2026.

1

Ideogram 4.5

Ideogram

Top‑tier at words that read correctly in the picture.

The board's typography specialist, now an editor too. 4.5 adds up to five reference images, masked edits and Precise Edit, which copies untouched pixels exactly so you can edit the same picture over many turns without it degrading.

Cost
API US$0.03 to US$0.10 an image from a prompt, US$0.008 to US$0.22 with references or Precise Edit
Where
KREA, OpenArt, Higgsfield · check in app: Magnific
Rights
Check terms.
Watch out
Photoreal and aesthetic depth below Midjourney and FLUX.
How to ask it[image type] of [main subject], [action or placement], [setting], [lighting], [style], with text that reads "[exact words]". To edit: [what changes], everything not mentioned stay exactly the same.

Put every word of on‑image copy inside double quotes.

Main idea first, in sentences.

Open Ideogram 4.5 in Call Sheet →
2

Nano Banana Pro / Imagen (Google)

Google

Best text rendering of the all‑rounders, and it handles other languages.

Precision + real‑world knowledge.

Cost
Mid (Pro). A faster/cheaper tier exists
Where
Magnific, Higgsfield, KREA, OpenArt, Runway
Rights
Google terms, plus a SynthID watermark.
Watch out
Less 'art‑directed' personality than Midjourney out of the box.
How to ask it[Subject] + [action] + [location] + [composition/shot] + [style & lighting]. Edits: [what to change] + [what to keep] + [refs]

Write full descriptive sentences, not keyword lists.

Put any on‑image text inside double quotes, spelled exactly as it should appear.

Open Nano Banana Pro / Imagen (Google) in Call Sheet →
3

GPT Image 2

OpenAI

Reliable text + conversational layout tweaks.

Conversational, instruction‑driven editing with excellent text.

Cost
Mid. Slower
Where
Magnific, Higgsfield, KREA, OpenArt, Runway
Rights
Check terms.
Watch out
Has a recognizable 'look'.
How to ask it[instruction] + [exact text in "quotes"] + [what to keep unchanged] + [style]

Write it as an instruction to a person, in plain sentences.

Quote any text that should appear in the image.

Open GPT Image 2 in Call Sheet →
4

Qwen Image 3

Alibaba (Qwen)

Dense layouts with text down to 10 pixels in 12 languages, per Qwen.

Qwen's third image model, built for dense layouts.

Cost
About US$0.04 at 1K and US$0.075 at 2K for the Pro tier, US$0.03 standard
Where
KREA · check in app: Magnific
Rights
Alibaba Cloud terms.
Watch out
Edits take only 1 to 3 images.
How to ask it[the image content, style and composition] + [every word to appear, in quotes] + [what to leave out, in the negative prompt]

Long, fully described prompts are what it is built for.

Quote every word that must appear.

Open Qwen Image 3 in Call Sheet →

Describe your own text in images job in Call Sheet and it ranks every model for it, then writes the prompt.

Open Call Sheet →

Head to head

More jobs