Сравнение аниме-пейзажей в четырех панелях

Wide horizontal social-media comparison graphic, 2:1 aspect ratio, set on a dark charcoal background with a black header band across the top and four tall vertical panels of equal width below, separated by thin black gutters. Header Text (top left): - Large white Japanese title: 「同じ構図・同じ画像AI（[GPT Image 2]）。プロンプトを“書いたAI”だけが違う4枚」 - Smaller muted gray subtitle directly beneath: 「変えたのは『プロンプトを書くClaudeのモデル階層』だけ（左＝最も安い → 右＝最上位）」 Panel Layout & Labels: Four vertical image panels arranged left to right, each with a dark label bar at the top displaying white text: 1. 「[haiku｜最も安い]」 2. 「[sonnet]」 3. 「[opus]」 4. 「[fable｜最上位]」 Panel Imagery & Art Direction (Makoto Shinkai-inspired anime cinematic landscape art, consistent core composition of a lone tiny human figure on a high grassy mountain ridge overlooking a vast valley and layered peaks): - Panel 1 ([haiku｜最も安い]): Fantasy sunset landscape with vivid orange and magenta skies, low glowing sun near the horizon, misty purple-blue distant mountains, foreground wildflowers, and a single tiny dark silhouette figure centered on the lower ridge. - Panel 2 ([sonnet]): Restrained natural lighting, expansive textured cumulus cloud mass across a blue-and-gold sky, soft atmospheric valley haze, and a solitary tiny figure standing at the edge of a mountain trail. - Panel 3 ([opus]): Dynamic cinematic composition with sweeping cloud waves, golden hour sun low on the right horizon illuminating a winding river in the valley below, crisp foreground grasses, and a lone tiny figure on a grassy crest. - Panel 4 ([fable｜最上位]): Vibrant anime film still lighting with turquoise-blue sky, towering cumulus clouds with radiant god rays breaking through, lush emerald green valley, and a small distinct figure of a girl in a white sundress and straw hat turning on the windswept slope. Style: High-end painterly anime background art, atmospheric perspective, cinematic bloom, clean UI presentation with strictly no extra watermarks, icons, or logos.