Image Generation API Pricing - April 2026
Per-image API costs for GPT Image 2, FLUX.2 Pro, Imagen 4, Ideogram v3, Stable Diffusion, and more - with price corrections and new additions for April 2026.
10 min read
TL;DR
- Cheapest per image: Stability AI SDXL at ~$0.003, or GPT Image 1 Mini (Low) at $0.005
- Best value at standard quality: FLUX.2 Pro at $0.03/1MP - unchanged, still competitive against GPT Image 2
- GPT Image 2 launched April 21 with token-based pricing: ~$0.006 (low), ~$0.053 (medium), ~$0.211 (high) at 1024x1024
- SD 3.5 Large was mispriced in March - it's $0.065, not $0.035; Recraft V4 Vector also corrected to $0.04
Quick Verdict
FLUX.2 Pro at $0.03/image stays the best bang-for-buck pick for standard production workloads. GPT Image 2 is the new quality ceiling but costs 1.8x more at medium quality (~$0.053 vs $0.034 for GPT Image 1.5 medium). Ideogram v3 Turbo at $0.03 earns a spot on the shortlist for any workflow where text rendering matters - it's cheaper than the previous Ideogram v2 API and the text accuracy improvement is real.
Two corrections from March: SD 3.5 Large was listed at ~$0.035 per image when the actual credit cost is $0.065 - nearly double. And Recraft V4 Vector is $0.04, not $0.08 as previously listed (the $0.08 tier is Recraft V4 Pro, a different product). Both errors are fixed below.
For deeper coverage of what these models can actually produce, see our best AI image generators roundup.
Normalized Pricing Table
All prices are per single image at standard resolution (~1024x1024, 1 megapixel). Sorted cheapest first. Prices verified April 27, 2026.
| Model | Provider | Price/Image | Resolution | Quality | API Type |
|---|---|---|---|---|---|
| SDXL | Stability AI | ~$0.003 | 1024x1024 | Standard | REST, credits |
| GPT Image 1 Mini (Low) | OpenAI | $0.005 | 1024x1024 | Low | REST |
| GPT Image 2 (Low) | OpenAI | ~$0.006 | 1024x1024 | Low | REST |
| GPT Image 1.5 (Low) | OpenAI | $0.009 | 1024x1024 | Low | REST |
| GPT Image 1 (Low) | OpenAI | $0.011 | 1024x1024 | Low | REST |
| FLUX.2 Klein 4B | BFL | $0.014 | 1MP | Standard | REST |
| FLUX.2 Klein 9B | BFL | $0.015 | 1MP | Standard | REST |
| GPT Image 1 Mini (Med) | OpenAI | $0.015 | 1024x1024 | Medium | REST |
| DALL-E 2 | OpenAI | $0.016 | 1024x1024 | Standard | REST, legacy |
| Imagen 4 Fast | $0.02 | 1024x1024 | Standard | REST/Vertex | |
| Stable Image Core | Stability AI | $0.03 | ~1024x1024 | Standard | REST, credits |
| FLUX.2 Pro | BFL | $0.03 | 1MP | Standard | REST |
| Ideogram v3 Turbo | Ideogram | $0.03 | 1024x1024 | Turbo | REST |
| Grok Aurora | xAI | $0.03 | 1024x1024 | Standard | REST |
| Gemini 2.5 Flash Image | ~$0.039 | 1024x1024 | Standard | REST (token-based) | |
| GPT Image 1.5 (Med) | OpenAI | $0.034 | 1024x1024 | Medium | REST |
| DALL-E 3 (Standard) | OpenAI | $0.04 | 1024x1024 | Standard | REST |
| FLUX 1.1 Pro | BFL | $0.04 | 1MP | Standard | REST |
| FLUX.1 Kontext [pro] | BFL | $0.04 | 1MP | Standard | REST |
| Recraft V4 | Recraft | $0.04 | 1024x1024 | Standard | REST |
| Recraft V4 Vector | Recraft | $0.04 | 1024x1024 | Standard | REST |
| GPT Image 1 Mini (High) | OpenAI | $0.052 | 1024x1024 | High | REST |
| GPT Image 2 (Med) | OpenAI | ~$0.053 | 1024x1024 | Medium | REST |
| FLUX.2 Flex | BFL | $0.06 | 1MP | Standard | REST |
| Imagen 4 | $0.06 | 1024x1024 | Ultra | REST/Vertex | |
| SD 3.5 Large | Stability AI | $0.065 | 1024x1024 | Standard | REST, credits |
| Gemini 3.1 Flash Image | ~$0.067 | 1024x1024 | Standard | REST (token-based) | |
| FLUX.2 Max | BFL | $0.07 | 1MP | Ultra | REST |
| Recraft V4 Pro | Recraft | $0.08 | 1024x1024 | Pro | REST |
| FLUX.1 Kontext [max] | BFL | $0.08 | 1MP | Ultra | REST |
| Stable Image Ultra | Stability AI | $0.08 | 1024x1024 | Ultra | REST, credits |
| Ideogram v3 Quality | Ideogram | $0.09 | 1024x1024 | Quality | REST |
| DALL-E 3 (HD) | OpenAI | $0.12 | 1024x1024 | HD | REST |
| GPT Image 1.5 (High) | OpenAI | $0.133 | 1024x1024 | High | REST |
| GPT Image 1 (High) | OpenAI | $0.167 | 1024x1024 | High | REST |
| GPT Image 2 (High) | OpenAI | ~$0.211 | 1024x1024 | High | REST |
| Recraft V4 Pro Vector | Recraft | $0.25 | 1024x1024 | Pro vector | REST |
What Changed Since March
Three significant corrections: SD 3.5 Large was listed at ~$0.035 when the Stability AI credit cost actually computes to $0.065. Recraft V4 Vector was listed at $0.08 - that's the V4 Pro price; the actual V4 Vector (raster output) is $0.04. FLUX.2 Flex climbed from $0.05 to $0.06/MP. Grok Imagine also rebranded to Grok Aurora with a price bump from $0.02 to $0.03.
Three new entries: GPT Image 2 (token-based, API opens May 2026), Ideogram v3 Turbo and Quality, and FLUX.2 Max. See the FLUX model comparison guide for a breakdown of when to use each BFL tier.
BFL's pricing calculator at bfl.ai/pricing - FLUX.2 charges per megapixel, so resolution inputs matter.
Source: bfl.ai
Subscription-Based Services
Midjourney still doesn't offer a first-party API. Their subscription plans give you GPU time, not per-call credits.
| Service | Plan | Monthly Cost | Est. Images/Month | Est. Cost/Image |
|---|---|---|---|---|
| Midjourney Basic | Subscription | $10 | ~200 | ~$0.05 |
| Midjourney Standard | Subscription | $30 | ~900 | ~$0.03 |
| Midjourney Pro | Subscription | $60 | ~1,800 | ~$0.03 |
| Midjourney Mega | Subscription | $120 | ~3,600 | ~$0.03 |
At Standard and above, Midjourney's blended cost with Relax mode lands around $0.03 per image. That's competitive with FLUX.2 Pro on price, but you're locked into their web interface. See our Midjourney v7 review for a full capability breakdown.
New This Month: GPT Image 2
OpenAI's official image generation pricing as of April 27, 2026 - all three models use per-token billing.
Source: developers.openai.com
GPT Image 2 launched for ChatGPT users on April 21, 2026. Developer API access is opening in early May 2026. It uses token-based pricing: $8/M image input tokens, $30/M image output tokens, and $5/M text input tokens.
In practice, that translates to roughly $0.006 at low quality, $0.053 at medium, and $0.211 at high for a standard 1024x1024 output. OpenAI has also added a 50% batch API discount - so production pipelines that tolerate async processing can cut those costs in half.
The model sits above GPT Image 1.5 in quality at medium tier ($0.053 vs $0.034), which is a meaningful price jump for a modest quality gain. Unless your output quality requirements specifically demand GPT Image 2, GPT Image 1.5 medium is still the smarter default until the API is generally available and independent quality comparisons are out.
Ideogram v3: Best Text Rendering at Budget Pricing
Ideogram 3.0 was released March 26, 2025 and has been quietly available through the API since then - it wasn't covered in the March pricing page. The Turbo tier at $0.03/image makes it competitive with FLUX.2 Pro. The Quality tier at $0.09 is expensive relative to alternatives.
What Ideogram actually delivers for that price: the best text rendering in any generation API at this price point, with roughly 90% accuracy on short phrases. FLUX.2 Pro, GPT Image 1.5, and Imagen 4 all struggle with legible text. If your workflow creates images with labels, signs, or interface text, Ideogram v3 Turbo is the only sub-$0.05 option worth considering.
The tradeoff is that Ideogram doesn't support editing workflows the way FLUX.1 Kontext does - it's a pure text-to-image model with optional style references.
Hidden Costs
Resolution Multipliers
FLUX.2 charges by megapixel. A 1MP image (1024x1024) costs the base rate; a 2MP image (1448x1448) costs roughly double. GPT Image models also scale with resolution - a 1536x1024 landscape from GPT Image 2 costs more than a 1024x1024 square at the same quality tier. Imagen 4 and Ideogram v3 are flat per-image regardless of common resolutions up to 1024x1024.
Quality Tiers Create 10-35x Price Ranges
OpenAI's three-tier system (low/medium/high) spans enormous cost ranges. GPT Image 2 goes from $0.006 to $0.211 per image - a 35x difference for the same model. Most benchmark comparisons report medium quality, but production customer-facing output usually needs high quality. Always test at your target quality level before committing to cost estimates.
Editing Surcharges
FLUX.2 charges 1.5x for editing (inpainting, outpainting) vs. generation. A FLUX.2 Pro edit is $0.045 vs. $0.03 for generation. FLUX.1 Kontext [pro] at $0.04 covers both generation and context-guided editing at the same price - useful for product variant workflows where you supply reference images. OpenAI charges the same rate for both generation and editing.
Batch Discounts
OpenAI offers 50% off image generation through their batch API. Google offers similar batch discounts on Vertex AI for Imagen calls. BFL's API has no published batch discount. For high-volume workloads (10,000+ images per day), the OpenAI batch pricing makes GPT Image 1 Mini at low quality truly competitive with Stable Image Core.
Rate Limits
OpenAI caps image generation at 50-500 images per minute depending on tier. Stability AI throttles under high load without published limits. BFL's API supports up to 100 concurrent requests. For burst workloads, rate limits matter more than per-image cost.
Free Tier Comparison
| Provider | Free Offering | Images Included | Limitations |
|---|---|---|---|
| Google (Gemini API) | Free tier | 1,500/day via Gemini Flash | Rate limited, lower resolution |
| OpenAI | $5 trial credits | ~1,000 (Mini Low) | Credits expire in 3 months |
| Stability AI | 25 free credits | ~8 images (SD 3.5 Large) | One-time |
| BFL (FLUX) | FLUX.2 Dev | Unlimited (self-hosted) | Requires own GPU |
| Midjourney | None | 0 | Trial discontinued |
Google wins the free tier. The Gemini Flash API free tier includes image generation at meaningful volume - the only provider offering that in production. For self-hosted options, FLUX.2 Dev is free to run locally if you have the hardware. See our guide on running open-source models locally for setup instructions.
Stable Image Core at $0.03 is Stability AI's answer to FLUX.2 Pro. It's faster and cheaper than SD 3.5 Large but lacks the photorealism of BFL's top tier.
Price History
Apr 2026 - OpenAI launched GPT Image 2 with token-based pricing. ChatGPT access started April 21; developer API access opening May 2026. Grok Imagine rebranded to Grok Aurora with a $0.01 price increase to $0.03/image.
Mar 2026 - BFL added FLUX.1 Kontext [pro] at $0.04 and [max] at $0.08 to its lineup. FLUX.2 Flex climbed from $0.05 to $0.06/MP.
Feb 2026 - OpenAI launched GPT Image 1.5 with 20-40% lower pricing than GPT Image 1 at equivalent quality tiers.
Jan 2026 - Google released Imagen 4 at $0.06/image (Ultra) and Imagen 4 Fast at $0.02, undercutting most mid-tier competitors. Gemini 2.5 Flash Image generation went live at ~$0.039 per 1024x1024 image.
Dec 2025 - BFL launched the FLUX.2 lineup. FLUX.2 Pro at $0.03 replaced FLUX 1.1 Pro ($0.04) as the default recommendation. FLUX.2 Max launched at $0.07 for the highest quality tier.
Nov 2025 - xAI introduced Grok Imagine at $0.02/image (now Aurora at $0.03).
Mar 2025 - Ideogram released v3.0 with Turbo ($0.03) and Quality ($0.09) tiers. First major competitor to beat Midjourney on text rendering accuracy.
The market has compressed from roughly $0.04-0.08 as the budget range in 2024 to $0.015-0.03 now. The quality gap between providers is also shrinking. FLUX.2 Pro and Imagen 4 Ultra both produce results that would have been premium-tier in early 2025.
FAQ
Which image generation API is cheapest?
Stability AI SDXL at ~$0.003/image is cheapest. Among major providers with actively maintained APIs, GPT Image 2 low-quality at ~$0.006 and Imagen 4 Fast at $0.02 are the budget picks.
What's the best value for production image generation?
FLUX.2 Pro at $0.03/1MP. Strong photorealism, no per-tier quality upcharge, and the same price covers both generation and editing via FLUX.1 Kontext.
What's GPT Image 2 and is it available now?
GPT Image 2 launched on April 21, 2026 for ChatGPT users. Developer API access opens in early May 2026. Pricing is token-based: roughly $0.053/image at medium quality, $0.211 at high.
Can I create images for free?
Google's Gemini API free tier includes image generation at 1,500 images/day via Gemini Flash. FLUX.2 Dev is free for local/self-hosted use. Midjourney discontinued its free trial.
How does Midjourney compare to API-based options?
At Standard tier ($30/month), Midjourney averages ~$0.03/image with Relax mode. Quality is competitive with FLUX.2 Pro. The constraint is zero official API access - everything runs through their web interface or unofficial wrappers.
Do higher resolutions cost more?
Yes, across all providers. BFL scales linearly by megapixel. OpenAI charges 50-100% more for landscape/portrait vs. square at the same quality tier. Ideogram v3 and Imagen 4 keep flat pricing up to 1024x1024.
Sources:
✓ Last verified April 27, 2026
