{
 "sym": "Im",
 "updated": "2026-08-06",
 "verdict": "gpt-image-2, if you can only hold one \u2014 since its April 21, 2026 launch it tops both public arenas by the widest margin any image model has ever held (~80 Elo over #2 on Arena, Aug 2026), follows layout-level instructions no rival matches, ships free inside ChatGPT, and OpenAI assigns you ownership of outputs. Midjourney V8.2 is the pick when the brand look is the product \u2014 it remains the aesthetic ceiling at a $10 entry. Nano Banana 2 wins when images run inside your product at volume ($0.045/image, batch $0.022). FLUX.2 wins when you need open weights, fine-tuning, and self-hosting. Ideogram 4.0 wins for typography-heavy brand assets at the cheapest production API prices.",
 "top5": [
  {
   "rank": 1,
   "name": "gpt-image-2",
   "maker": "OpenAI",
   "url": "https://chatgpt.com",
   "docs": "https://developers.openai.com/api/docs/guides/image-generation",
   "pricing": "Free in ChatGPT (all tiers, since Apr 21, 2026) \u00b7 API token-priced: $5/1M text in, $8/1M image in, $30/1M out (batch half) \u00b7 gpt-image-1.5 out $10/1M \u00b7 gpt-image-1-mini out $8/1M",
   "best_for": "The one-tool default: brand-grade visuals from plain-English direction \u2014 precise text, layouts, multi-turn edits, reference images \u2014 in the chat app you already pay for.",
   "why": "The first image model to lead both public leaderboards decisively: 1380 Elo on Arena vs 1301 for #2 Reve (Aug 4, 2026, 5.86M votes) and 1339 on Artificial Analysis. Launched in ChatGPT as \"Images 2.0\" on Apr 21, 2026 with images-with-thinking planning mode; the API adds mask editing, streaming, and multi-turn reference workflows. OpenAI's terms assign output ownership to you \u2014 the cleanest commercial-rights posture of the closed labs.",
   "watch": "A recognizable house style \u2014 arena wins measure preference, not brand distinctiveness, and Midjourney users can spot gpt-image output instantly. Generation is slower than Nano Banana 2, token-priced API costs are hard to predict per image, and outputs are not unique \u2014 other users can get near-identical results.",
   "evidence": [
    {
     "stat": "#1 Arena text-to-image: 1380 Elo vs 1301 for #2 (Aug 4, 2026, 5,859,509 votes)",
     "src": "https://arena.ai/leaderboard/text-to-image"
    },
    {
     "stat": "#1 Artificial Analysis: GPT Image 2 (high) 1339 Elo (Aug 2026)",
     "src": "https://artificialanalysis.ai/text-to-image/arena/leaderboard-text"
    },
    {
     "stat": "ChatGPT Images 2.0 launched to all subscription tiers Apr 21, 2026; \"images with thinking\" on paid plans",
     "src": "https://help.openai.com/en/articles/6825453-chatgpt-release-notes"
    }
   ],
   "tile_note": "the instruction-following default"
  },
  {
   "rank": 2,
   "name": "Midjourney",
   "maker": "Midjourney, Inc. (bootstrapped)",
   "url": "https://www.midjourney.com",
   "docs": "https://docs.midjourney.com",
   "pricing": "Basic $10/mo \u00b7 Standard $30 (unlimited relax) \u00b7 Pro $60 (stealth) \u00b7 Mega $120; annual \u221220% \u00b7 companies >$1M revenue must use Pro/Mega",
   "best_for": "Founders for whom the visual identity is the moat \u2014 art direction, personalization profiles trained on your ratings, and a look nothing else matches.",
   "why": "Still the aesthetic ceiling: V8 alpha (Mar 17, 2026) through V8.2 (Jul 24, 2026) pushed \"more creative, bold, sophisticated\" output plus personalization that learns your taste from rating history \u2014 the closest thing to a house style engine. Commercial terms are simple: any subscriber can use images commercially. The company stays independent and profitable enough to make its first acquisition (Co-Star, Jul 23, 2026) and tease its own hardware (Jun 16, 2026).",
   "watch": "Refuses public benchmarks, so no independent quality ranking exists; still no public API, so it cannot sit in an automated pipeline; the Disney/Universal copyright suit (filed Jun 2025) is unresolved; and prompt-precision and text rendering trail gpt-image-2 and Ideogram.",
   "evidence": [
    {
     "stat": "V8.2 released Jul 24, 2026 \u2014 aesthetics, quality, and personalization upgrades",
     "src": "https://updates.midjourney.com/version-8-2/"
    },
    {
     "stat": "Plans $10\u2013120/mo; commercial use for all subscribers, Pro/Mega required above $1M annual revenue",
     "src": "https://docs.midjourney.com/hc/en-us/articles/27870484040333-Version"
    },
    {
     "stat": "First acquisition (Co-Star) Jul 23, 2026; hardware announcement livestream Jun 16, 2026",
     "src": "https://updates.midjourney.com/midjourneys-first-acquisition/"
    }
   ],
   "tile_note": "the aesthetic ceiling"
  },
  {
   "rank": 3,
   "name": "Nano Banana 2 (Gemini 3.1 Flash Image)",
   "maker": "Google DeepMind",
   "url": "https://gemini.google.com",
   "docs": "https://ai.google.dev/gemini-api/docs/image-generation",
   "pricing": "Free in Gemini app \u00b7 API ~$0.045/image (0.5K) \u00b7 $0.067 (1K) \u00b7 batch half price \u00b7 Nano Banana Pro (Gemini 3 Pro Image): $0.134/1\u20132K, $0.24/4K",
   "best_for": "Product-embedded generation and conversational editing at volume \u2014 the best price-performance of any frontier image API, with 4K available on Pro.",
   "why": "The viral editing king industrialized: the Nano Banana line now spans Lite, 2, and Pro (2K/4K) tiers, all top-10 on Arena (1262/1251/1245 Elo, Aug 4, 2026), at prices an order of magnitude below closed rivals \u2014 $0.045 standard, $0.022 batch. Distribution is unmatched: free in the Gemini app, inside Workspace, and licensed into Adobe Firefly as a partner model.",
   "watch": "Aesthetics and art direction trail Midjourney and gpt-image-2 \u2014 it wins on speed, edits, and price, not on look. SynthID watermarking is non-optional, and person-generation policies are stricter than rivals'.",
   "evidence": [
    {
     "stat": "gemini-3.1-flash-image 1262 Elo, #6 on Arena; three Nano Banana models in the top 10 (Aug 4, 2026)",
     "src": "https://arena.ai/leaderboard/text-to-image"
    },
    {
     "stat": "API pricing: $0.045/0.5K image, $0.067/1K; batch $0.022/$0.034; Pro $0.134\u20130.24 (Aug 2026)",
     "src": "https://ai.google.dev/gemini-api/docs/pricing"
    },
    {
     "stat": "Licensed into Adobe Firefly as partner model alongside GPT Image 1.5 and FLUX.2 (Aug 2026)",
     "src": "https://www.adobe.com/products/firefly.html"
    }
   ],
   "tile_note": "cheap, fast, everywhere"
  },
  {
   "rank": 4,
   "name": "FLUX.2",
   "maker": "Black Forest Labs",
   "url": "https://bfl.ai",
   "docs": "https://docs.bfl.ai",
   "pricing": "API: FLUX.2 [klein] from $0.014 \u00b7 [pro] from $0.03 \u00b7 [max] from $0.07 \u00b7 FLUX1.1 [pro] $0.04 \u00b7 self-host commercial licenses from Builder tier (10K img/mo); fine-tuning/LoRA rights included",
   "best_for": "Teams that need the model in their pipeline \u2014 open weights, LoRA fine-tuning on brand assets, self-hosting, and no dependency on a closed lab's roadmap.",
   "why": "The open-weights standard-bearer now that Stability has pivoted to enterprise services: the FLUX.2 family spans Apache-licensed klein (4B/9B) through frontier [max], with fine-tuning and LoRA rights in every commercial tier \u2014 the only top-tier lab where you can own the deployment. A $300M Series B and Martin Scorsese as advisor signal staying power; FLUX 3 (multimodal image/video/audio) is the successor line.",
   "watch": "Quality has slipped relative to the closed frontier \u2014 FLUX.2 [max] sits at 1196 Elo, #15 on Artificial Analysis (Aug 2026). Licensing is genuinely confusing (non-commercial dev weights vs paid self-host tiers vs API), and the FLUX 3 push into video/audio risks diluting image focus.",
   "evidence": [
    {
     "stat": "API prices: klein $0.014\u20130.015, pro $0.03, max $0.07 per image (Aug 2026)",
     "src": "https://docs.bfl.ai/quick_start/pricing"
    },
    {
     "stat": "$300M Series B closed; Martin Scorsese joins as advisor; FLUX 3 multimodal launched (2026)",
     "src": "https://bfl.ai/"
    },
    {
     "stat": "FLUX.2 [max] 1196 Elo, #15 on Artificial Analysis (Aug 2026) \u2014 behind the closed frontier",
     "src": "https://artificialanalysis.ai/text-to-image/arena/leaderboard-text"
    }
   ],
   "tile_note": "open weights, your pipeline"
  },
  {
   "rank": 5,
   "name": "Ideogram",
   "maker": "Ideogram AI",
   "url": "https://ideogram.ai",
   "docs": "https://developer.ideogram.ai",
   "pricing": "App freemium \u00b7 API per-image, no subscription: 4.0 Turbo $0.03 \u00b7 4.0 Default $0.06 \u00b7 4.0 Quality $0.10 \u00b7 3.0 from $0.03 \u00b7 character-reference variants $0.10\u20130.20",
   "best_for": "Text-in-image work \u2014 logos, posters, packaging, ad creative, social graphics \u2014 where typography must render exactly, at the lowest production API prices.",
   "why": "Ideogram 4.0 keeps the typography crown: \"crystal-clear type, prompt fidelity, reliable editing\" with mask editing, background replacement, style presets, and custom model training via API \u2014 a full production-graphics toolchain priced at $0.03\u20130.10 per image with volume discounts and no subscription required. Founded by the Imagen authors ($16.5M a16z/Index seed 2023, $80M Feb 2024), it stays the specialist the giants haven't displaced.",
   "watch": "General realism trails the frontier \u2014 ideogram-4.0-quality ranks #14 on Arena (1205 Elo, Aug 2026) \u2014 and no new funding has been announced since Feb 2024 while every hyperscaler now ships text-capable image models. Its edge is narrowing from \"only one that can spell\" to \"best design workflow.\"",
   "evidence": [
    {
     "stat": "Ideogram 4.0 API: $0.03 Turbo / $0.06 Default / $0.10 Quality per image (Aug 2026)",
     "src": "https://ideogram.ai/api-pricing"
    },
    {
     "stat": "ideogram-4.0-quality: 1205 Elo, #14 on Arena (Aug 4, 2026)",
     "src": "https://arena.ai/leaderboard/text-to-image"
    },
    {
     "stat": "$80M raised Feb 2024 after 1.0; founded 2022 by Imagen alumni (Norouzi, Chan, Saharia, Ho)",
     "src": "https://en.wikipedia.org/wiki/Ideogram_(text-to-image_model)"
    }
   ],
   "tile_note": "text that renders right"
  }
 ],
 "matrix": {
  "cols": [
   "Latest model (Aug '26)",
   "Best at",
   "Entry price",
   "API $/image",
   "Open weights",
   "Commercial rights",
   "Arena Elo (Aug '26)"
  ],
  "rows": [
   [
    "gpt-image-2",
    "gpt-image-2 (Apr 2026)",
    "Instruction following, edits",
    "Free in ChatGPT",
    "Token-priced ($30/1M out)",
    "No",
    "You own outputs",
    "1380 (#1)"
   ],
   [
    "Midjourney",
    "V8.2 (Jul 2026)",
    "Aesthetics, house style",
    "$10/mo",
    "No public API",
    "No",
    "All subscribers; >$1M rev needs Pro/Mega",
    "Not benchmarked (declines arenas)"
   ],
   [
    "Nano Banana 2 (Gemini 3.1 Flash Image)",
    "Gemini 3.1 Flash Image (2026)",
    "Speed, price, conversational edits",
    "Free in Gemini",
    "$0.045\u20130.067; batch $0.022",
    "No",
    "Commercial use; SynthID watermark",
    "1262 (#6)"
   ],
   [
    "FLUX.2",
    "FLUX.2 [max]; FLUX 3 multimodal",
    "Self-host, fine-tuning, control",
    "Pay-as-you-go",
    "$0.014\u20130.07",
    "Yes (klein; dev non-commercial)",
    "Paid self-host licenses w/ LoRA rights",
    "1196 AA (#15)"
   ],
   [
    "Ideogram",
    "4.0 (2026)",
    "Typography, design assets",
    "Free app tier",
    "$0.03\u20130.10",
    "No",
    "Commercial use on API/paid",
    "1205 (#14)"
   ],
   [
    "Reve",
    "Reve 2.1 (Jul 9, 2026)",
    "4K quality + canvas editor",
    "Free app tier",
    "API since Jul 14, 2026",
    "No",
    "Commercial on paid",
    "1301 (#2)"
   ],
   [
    "Recraft",
    "V4.1 (2026)",
    "Brand kits, vector/SVG, mockups",
    "Free (non-commercial)",
    "Per-image API",
    "No",
    "Paid = full ownership; free images owned by Recraft",
    "1205 AA (#12)"
   ],
   [
    "Seedream",
    "5.0 Pro (2026)",
    "Photoreal quality per dollar",
    "Via API platforms",
    "Low (platform-dependent)",
    "No",
    "Per platform terms",
    "1257 (#7)"
   ]
  ]
 },
 "rules": [
  {
   "if": "You need one default and already live in ChatGPT",
   "then": "gpt-image-2 \u2014 the #1 model on both arenas at zero marginal cost, and you own the outputs outright."
  },
  {
   "if": "The brand look is the product and you'll invest taste, not just prompts",
   "then": "Midjourney at $10\u201330/mo \u2014 rate images to train a personalization profile, and treat your prompt library like a brand book. Upgrade to Pro past $1M revenue; it's in the terms."
  },
  {
   "if": "Images render inside your product, or you generate thousands per month",
   "then": "Nano Banana 2 API at $0.045/image ($0.022 batch) \u2014 an order of magnitude cheaper than closed rivals; step up to Nano Banana Pro only when you need 2K/4K."
  },
  {
   "if": "You need fine-tuning on brand assets, self-hosting, or independence from any lab's roadmap",
   "then": "FLUX.2 \u2014 klein weights to prototype, a Builder-tier license ($10K img/mo, LoRA rights) to ship."
  },
  {
   "if": "The deliverable is words-on-image \u2014 logos, packaging, ads, posters \u2014 or you need contractual full ownership",
   "then": "Ideogram 4.0 API ($0.03\u20130.10/image) for typography; Recraft on a paid plan when the client demands you own every pixel."
  }
 ],
 "field": [
  {
   "name": "Reve",
   "maker": "Reve AI",
   "note": "The near-miss for the top 5: Reve 2.1 (Jul 9, 2026) is #2 on both arenas (1301/1299 Elo) with a genuinely novel canvas editor (Live Layers, Reframe, sketch-to-image); API launched Jul 14, 2026 \u2014 young, but the fastest riser of 2026",
   "url": "https://reve.com",
   "oss": false,
   "entry": "free tier",
   "status": "active"
  },
  {
   "name": "Recraft",
   "maker": "Recraft",
   "note": "Design-first generator (V4.1): vector/SVG output, brand styles, mockups; sharpest ownership terms in the category \u2014 paid plans grant full ownership, free-plan images belong to Recraft",
   "url": "https://www.recraft.ai",
   "oss": false,
   "entry": "free (non-commercial)",
   "status": "active"
  },
  {
   "name": "Seedream",
   "maker": "ByteDance",
   "note": "Seedream 5.0 Pro at 1257 Elo (#7 Arena, Aug 2026) \u2014 frontier photorealism per dollar via BytePlus/fal/partner APIs; sibling Seedance owns element Vi",
   "url": "https://seed.bytedance.com",
   "oss": false,
   "entry": "API via platforms",
   "status": "active"
  },
  {
   "name": "Muse Image",
   "maker": "Meta",
   "note": "Meta's frontier entry (Jul 7, 2026), launched with Muse Video \u2014 instruction-faithful, multi-reference composition, 'draws on Instagram for social context'; #3 on Arena at 1282",
   "url": "https://ai.meta.com/blog/",
   "oss": false,
   "entry": "free in Meta apps",
   "status": "active"
  },
  {
   "name": "MAI-Image-2.5",
   "maker": "Microsoft AI",
   "note": "Microsoft's in-house model, #8 on Arena (1255) \u2014 powers Copilot/Bing image generation as Microsoft weans off OpenAI models",
   "url": "https://microsoft.ai",
   "oss": false,
   "entry": "free in Copilot",
   "status": "active"
  },
  {
   "name": "Grok Imagine",
   "maker": "SpaceX (xAI, acquired Feb 2, 2026)",
   "note": "Image+video generator with Quality Mode API (May 6, 2026) \u2014 stronger realism and text rendering; 1228 Elo on Arena; minimal content filtering remains its differentiator and its risk",
   "url": "https://x.ai",
   "oss": false,
   "entry": "X Premium bundle",
   "status": "active"
  },
  {
   "name": "Qwen-Image",
   "maker": "Alibaba",
   "note": "qwen-image-3.0-pro is #5 on Arena (1263) \u2014 the strongest open-weights-lineage lab; earlier Qwen-Image weights were Apache-2.0, 3.0-pro is API-first",
   "url": "https://qwen.ai",
   "oss": true,
   "entry": "free tier + API",
   "status": "active"
  },
  {
   "name": "HiDream",
   "maker": "HiDream (Vivago)",
   "note": "HiDream-O1-Image-1.5 at 1245 Elo (#7 Artificial Analysis) \u2014 the open-weights dark horse from China, MIT-licensed earlier releases",
   "url": "https://hidream.ai",
   "oss": true,
   "entry": "free + tokens",
   "status": "active"
  },
  {
   "name": "Adobe Firefly",
   "maker": "Adobe",
   "note": "Pivoted from model lab to model hub: Firefly Image 5 plus partner models (GPT Image 1.5, FLUX.2 [pro], Nano Banana 2) under one $9.99\u201349.99/mo roof with 'commercially safe' indemnified training data \u2014 the enterprise-procurement answer",
   "url": "https://www.adobe.com/products/firefly.html",
   "oss": false,
   "entry": "free \u00b7 $9.99/mo",
   "status": "active"
  },
  {
   "name": "Stable Diffusion / Stability AI",
   "maker": "Stability AI",
   "note": "The category's founding open-source lab, now an enterprise services company (Brand Studio, custom models for EA/UMG/HubSpot; Stable Audio 3.0 is the 2026 flagship). No frontier image model since SD 3.5 (Oct 2024); UK High Court largely cleared it in the Getty case (Nov 2025). The open-weights torch passed to BFL",
   "url": "https://stability.ai",
   "oss": true,
   "entry": "free weights \u00b7 enterprise",
   "status": "fading"
  },
  {
   "name": "Leonardo.ai",
   "maker": "Canva (acquired Jul 2024)",
   "note": "Prosumer generation platform (Lucid model line) powering Canva's AI features; increasingly a Canva feature rather than a standalone frontier lab \u2014 Canva itself belongs to element De",
   "url": "https://leonardo.ai",
   "oss": false,
   "entry": "free \u00b7 ~$12/mo",
   "status": "acquired"
  },
  {
   "name": "Krea",
   "maker": "Krea AI",
   "note": "Best-in-class multi-model workspace (Krea 2 in-house model plus FLUX, Ideogram, Runway, Z Image; realtime canvas, upscaling, LoRA training) \u2014 $5.25\u201363/mo, commercial license on all paid plans",
   "url": "https://www.krea.ai",
   "oss": false,
   "entry": "free \u00b7 $5.25/mo",
   "status": "active"
  },
  {
   "name": "Freepik AI Suite",
   "maker": "Freepik",
   "note": "Stock giant turned aggregator \u2014 Mystic model plus third-party frontier models, bundled with stock library; owns Magnific",
   "url": "https://www.freepik.com/ai",
   "oss": false,
   "entry": "free \u00b7 ~$12/mo",
   "status": "active"
  },
  {
   "name": "Magnific",
   "maker": "Freepik (acquired 2024)",
   "note": "The upscaler/enhancer standard for making AI images print-ready; now inside Freepik's suite",
   "url": "https://magnific.ai",
   "oss": false,
   "entry": "~$39/mo",
   "status": "acquired"
  },
  {
   "name": "Civitai",
   "maker": "Civit AI, Inc.",
   "note": "The SD/FLUX LoRA community hub \u2014 still the largest fine-tune marketplace, but bruised by 2025 payment-processor purges and content-policy turmoil",
   "url": "https://civitai.com",
   "oss": true,
   "entry": "free",
   "status": "active"
  },
  {
   "name": "Luma Photon",
   "maker": "Luma AI",
   "note": "Capable, cheap image line, but Luma's center of gravity is video (Ray models \u2192 element Vi); Photon has not kept frontier pace in 2026",
   "url": "https://lumalabs.ai",
   "oss": false,
   "entry": "free tier",
   "status": "fading"
  },
  {
   "name": "Playground",
   "maker": "Playground AI",
   "note": "Once a leading SD interface (Playground v3), pivoted to consumer logo/design tooling in 2025 \u2014 no frontier model since",
   "url": "https://playground.com",
   "oss": false,
   "entry": "free tier",
   "status": "fading"
  },
  {
   "name": "DALL-E 3",
   "maker": "OpenAI",
   "note": "The name that defined the category, retired in practice \u2014 superseded by the autoregressive gpt-image line (gpt-image-1 Apr 2025 \u2192 1.5 Dec 2025 \u2192 2 Apr 2026)",
   "url": "https://openai.com/index/dall-e-3/",
   "oss": false,
   "entry": "\u2014",
   "status": "renamed"
  },
  {
   "name": "Amazon Nova Canvas",
   "maker": "AWS",
   "note": "Bedrock's in-house image model \u2014 enterprise checkbox, not a creative contender",
   "url": "https://aws.amazon.com/ai/generative-ai/nova/",
   "oss": false,
   "entry": "per-image via Bedrock",
   "status": "active"
  },
  {
   "name": "Getty Images Generative AI",
   "maker": "Getty Images",
   "note": "Fully-indemnified, licensed-training generator for legal-conservative enterprises; merged entity with Shutterstock consolidates the 'commercially safe stock' lane",
   "url": "https://www.gettyimages.com/ai/generation/overview",
   "oss": false,
   "entry": "per-generation packs",
   "status": "active"
  },
  {
   "name": "Lexica",
   "maker": "Lexica",
   "note": "Early SD search engine + Aperture model; effectively dormant since 2024",
   "url": "https://lexica.art",
   "oss": false,
   "entry": "\u2014",
   "status": "dead"
  },
  {
   "name": "NightCafe",
   "maker": "NightCafe Studio",
   "note": "Multi-model consumer community app \u2014 hobbyist lane, not startup-grade tooling",
   "url": "https://creator.nightcafe.studio",
   "oss": false,
   "entry": "free \u00b7 credits",
   "status": "active"
  }
 ],
 "signals": [
  {
   "fact": "gpt-image-2 leads Arena by ~80 Elo (1380 vs 1301, Aug 4, 2026, 5.86M votes) \u2014 the most decisive #1 the image category has ever had, confirmed independently by Artificial Analysis",
   "src": "https://arena.ai/leaderboard/text-to-image"
  },
  {
   "fact": "Every hyperscaler now ships a frontier image model: Meta launched Muse Image Jul 7, 2026; Microsoft's MAI-Image-2.5 is top-10; Google runs three Nano Banana tiers; SpaceX acquired xAI (Feb 2, 2026) and its Grok Imagine",
   "src": "https://ai.meta.com/blog/"
  },
  {
   "fact": "Black Forest Labs closed a $300M Series B and added Martin Scorsese as advisor; FLUX 3 extends the line to multimodal image/video/audio (2026)",
   "src": "https://bfl.ai/"
  },
  {
   "fact": "Midjourney stays bootstrapped and expansionist: first acquisition (Co-Star, Jul 23, 2026), hardware announcement livestream (Jun 16, 2026), V8\u2192V8.2 shipped Mar\u2013Jul 2026",
   "src": "https://updates.midjourney.com/"
  },
  {
   "fact": "Production API prices collapsed to $0.014\u20130.10 per image (FLUX.2 klein to Ideogram 4.0 Quality; Google batch $0.022) \u2014 image generation is now a rounding error in startup budgets",
   "src": "https://docs.bfl.ai/quick_start/pricing"
  },
  {
   "fact": "Stability AI \u2014 the company that started the open-source category \u2014 pivoted to enterprise brand services (EA, UMG, HubSpot) with no frontier image model since SD 3.5 (Oct 2024); the UK High Court largely cleared it in Getty's landmark training-data case (Nov 2025)",
   "src": "https://stability.ai/"
  }
 ],
 "notes": "Ranking criteria: independent arena standing (Arena + Artificial Analysis cross-checked Aug 2026), verified pricing, commercial-rights clarity, and fit for startup brand work \u2014 editorial, no affiliate consideration. Canon-pick change: Midjourney held the #1 tile through 2026.Q2; gpt-image-2's dual-leaderboard dominance since April and its clean ownership terms move it to #1 \u2014 Midjourney's absence from every public benchmark (it declines arena participation, so 'aesthetic ceiling' rests on practitioner consensus, not measurable Elo) is itself part of the case. Conflicts resolved: Arena and Artificial Analysis order the mid-board differently (e.g. Seedream 5.0 Pro #7 vs #8, MAI-Image-2.5 #8 vs #3); we cite each stat from the named board and rank on the aggregate picture. Weak/unverified claims flagged: Midjourney 'no public API' is practitioner consensus, not a vendor statement \u2014 re-check; Ideogram's consumer subscription prices and free-tier commercial terms could not be verified on-site (API prices are verified) \u2014 the Wikipedia entry is stale (stops at 3.0, Mar 2025, while ideogram-4.0 is live on Arena and in developer docs); Recraft's exact plan prices were not retrievable (ownership terms verified verbatim); Leonardo pricing page blocked our fetch (403); FLUX.2 [klein]'s exact open license and Seedream 5.0 per-image price were not verified on primary pages. gpt-image per-image cost is token-denominated \u2014 comparisons with flat per-image rivals are approximate. No credible 2026 market-size figure for image generation specifically survived source checks (Grand View's report page is 'being developed'), so none is cited. Adjacent elements: video models (Veo, Seedance, Runway, Grok Imagine Video, FLUX 3 video) \u2192 Vi; Canva/Figma design suites \u2192 De; deck illustration \u2192 Dk; ad-creative volume tools (AdCreative) \u2192 Pd; avatars \u2192 Av; upscalers (Magnific) noted here as workflow adjuncts.",
 "sources": [
  "https://arena.ai/leaderboard/text-to-image",
  "https://artificialanalysis.ai/text-to-image/arena/leaderboard-text",
  "https://help.openai.com/en/articles/6825453-chatgpt-release-notes",
  "https://developers.openai.com/api/docs/pricing",
  "https://developers.openai.com/api/docs/guides/image-generation",
  "https://openai.com/policies/terms-of-use/",
  "https://updates.midjourney.com/version-8-2/",
  "https://updates.midjourney.com/midjourneys-first-acquisition/",
  "https://docs.midjourney.com/hc/en-us/articles/27870484040333-Version",
  "https://ai.google.dev/gemini-api/docs/pricing",
  "https://ai.google.dev/gemini-api/docs/image-generation",
  "https://bfl.ai/",
  "https://docs.bfl.ai/quick_start/pricing",
  "https://ideogram.ai/api-pricing",
  "https://developer.ideogram.ai/",
  "https://en.wikipedia.org/wiki/Ideogram_(text-to-image_model)",
  "https://blog.reve.com/",
  "https://www.recraft.ai/pricing",
  "https://www.adobe.com/products/firefly.html",
  "https://stability.ai/",
  "https://en.wikipedia.org/wiki/Stability_AI",
  "https://x.ai/news",
  "https://ai.meta.com/blog/",
  "https://www.krea.ai/pricing"
 ],
 "element": {
  "number": 19,
  "name": "Image",
  "group": "Creative & Design",
  "essential": true,
  "edition": "v2026.Q3",
  "revision": "r7",
  "license": "CC BY 4.0 \u2014 cite elems.ai",
  "url": "https://elems.ai/e/im.html"
 }
}