docs: add 4 tech-style illustrations to Opus 4.7 blog post

Generated via gemini-3-pro-image-preview at 16:9. Matches AgentHunter
brand palette (deep navy + electric cyan). One image per major beat:
hero, rankings dashboard, speed race, cost scale.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: iamzifei

docs/blog/devto-opus-47-first-look.md

@@ -11,6 +11,8 @@ cover_image:
 
 Anthropic released **Claude Opus 4.7** on April 17, 2026. I ran it through the same 10-task evaluation I used for Opus 4.6, Sonnet 4.6, and Haiku 4.5 — this time with real token tracking so I could report dollar cost, not just pass rate.
 
+![Four Claude models on a benchmarking apparatus — Opus 4.7, Opus 4.6, Sonnet 4.6, Haiku 4.5](imgs/01-hero-four-claudes.png)
+
 ## TL;DR
 
 | Model | Tasks Passed | Avg Time | Total Cost | Cost / Task |
@@ -20,6 +22,8 @@ Anthropic released **Claude Opus 4.7** on April 17, 2026. I ran it through the s
 | Claude Sonnet 4.6 | 10/10 | 9.8s | $0.110 | $0.011 |
 | Claude Haiku 4.5 | 8/10 | 4.6s | $0.030 | $0.003 |
 
+![Ranking dashboard showing pass rate and total cost per model](imgs/02-rankings-dashboard.png)
+
 **Opus 4.7 is the new accuracy king and it's also faster than 4.6.** It costs ~27% more than 4.6 in total ($0.56 vs $0.44) but finishes tasks 14% faster on average. If you were using Opus 4.6, there's no reason not to upgrade.
 
 **Sonnet 4.6 is the sleeper.** Perfect 10/10 accuracy at **1/5 the cost** of Opus 4.7. Unless you specifically need the extra edge Opus brings on adversarial tasks, Sonnet is the right default for most production agent work.
@@ -52,12 +56,16 @@ Five coding tasks, five writing tasks. All graded by an independent LLM judge ag
 
 ### 1. Opus 4.7 is faster than 4.6, not slower
 
+![Speed comparison: Opus 4.7 trail 1.16x faster than Opus 4.6](imgs/03-speed-race.png)
+
 This is the surprise. Model version bumps usually trade off speed for capability — bigger model, longer generations. Opus 4.7 is the opposite: **8.4s average vs 4.6's 9.8s**, a 14% improvement. On the README task specifically (the longest task in the suite), 4.7 finished in 20.6s vs 4.6's 22.7s.
 
 Same pass rate, less latency, ~27% more cost. For interactive agent workloads where latency matters, the upgrade is worth it.
 
 ### 2. Sonnet 4.6 is the cost-adjusted winner
 
+![Balance scale: Sonnet at $0.11 vs Opus 4.7 at $0.56, both scoring 10/10](imgs/04-cost-scale.png)
+
 Sonnet 4.6 matches Opus 4.7's 10/10 accuracy on this suite at **$0.11 total vs $0.56** — **5× cheaper**. The gap between Sonnet and Opus used to be "Sonnet is fine if you're okay with 90% accuracy." As of this benchmark, there's no accuracy gap on these 10 tasks.
 
 Where Opus still earns its premium: tasks in the suite don't include adversarial inputs, long-context reasoning, or multi-step planning. For narrow, well-specified tasks like these, Sonnet is enough.

docs/blog/imgs/01-hero-four-claudes.png

@@ -0,0 +1,41 @@
+# Illustration Plan — Opus 4.7 First Look
+
+**Article**: docs/blog/devto-opus-47-first-look.md
+**Style**: `tech` (matches AgentHunter brand — dark navy #0a1628 + electric cyan #00d4ff)
+**Illustration Count**: 4 images
+
+---
+
+## Illustration 1 — Hero: "The Four Claudes"
+
+**Insert Position**: After opening paragraph, before TL;DR
+**Purpose**: Set atmosphere — position the story as a four-way head-to-head
+**Visual Content**: Four abstract AI "agent" orbs/nodes lined up on a dark blue technical grid, each labeled with its model name. The leftmost orb (Opus 4.7) glows brighter than the others. Connecting lines/wires suggest benchmarking apparatus. Crosshair/reticle overlay echoes the AgentHunter logo.
+**Filename**: 01-hero-four-claudes.png
+
+---
+
+## Illustration 2 — Rankings Dashboard
+
+**Insert Position**: Right after the TL;DR table
+**Purpose**: Visualize the benchmark scoreboard in a way that makes the numbers emotionally legible
+**Visual Content**: Stylized dashboard showing four horizontal bars (one per model) with pass-rate fill + cost overlay. Electric cyan glow on the bars for perfect scorers; Haiku's bar is shorter (8/10). Monospace numeric labels. Dark gradient background with subtle grid.
+**Filename**: 02-rankings-dashboard.png
+
+---
+
+## Illustration 3 — Speed: 4.7 beats 4.6
+
+**Insert Position**: In Key Findings #1 ("Opus 4.7 is faster than 4.6, not slower")
+**Purpose**: Convey the counterintuitive finding that the newer model is faster
+**Visual Content**: Two parallel speed tracks/lanes with abstract light trails. The Opus 4.7 trail is noticeably ahead of Opus 4.6, with a subtle "1.16×" badge. Stopwatch icon in the corner. Cyan-on-dark-blue. Motion blur effect suggests speed.
+**Filename**: 03-speed-race.png
+
+---
+
+## Illustration 4 — Cost: Sonnet is the Sleeper
+
+**Insert Position**: In Key Findings #2 ("Sonnet 4.6 is the cost-adjusted winner")
+**Purpose**: Highlight the 5× cost gap between Sonnet and Opus 4.7 at equal accuracy
+**Visual Content**: Balanced scale / pan balance weighing two stacked coin towers. The Sonnet side (smaller tower, $0.11) is labeled with a star/check badge. The Opus 4.7 side has a much taller tower ($0.56). Both sides have "10/10" score tags to emphasize equal accuracy. A glowing "5× cheaper" marker between them. Deep blue background, cyan accents on coins.
+**Filename**: 04-cost-scale.png

docs/blog/imgs/02-rankings-dashboard.png

@@ -0,0 +1 @@
+A wide cinematic tech illustration, 16:9 aspect ratio, depicting four abstract AI agent "orbs" arranged in a horizontal line on a dark navy blue technical grid background. Each orb is a glowing sphere made of concentric cyan rings with faint wireframe detail. From left to right they are labeled in thin monospace type: "OPUS 4.7", "OPUS 4.6", "SONNET 4.6", "HAIKU 4.5". The leftmost orb (OPUS 4.7) glows significantly brighter, with a halo of electric cyan light and a subtle "NEW" badge above it. Thin luminous measurement lines connect the orbs at the base, suggesting a benchmarking apparatus. A faint crosshair reticle with corner brackets is overlaid in the upper-right corner (echoing a scope/targeting aesthetic). Color palette: deep navy #0a1628 background, electric cyan #00d4ff primary accent, soft violet purple #6B46C1 secondary, white #FFFFFF for labels only. Style: modern minimalist tech infographic, clean geometric shapes, no human figures, no realistic faces. Dark atmospheric lighting. Professional editorial illustration quality. Negative space emphasized. 16:9 ratio.

docs/blog/imgs/03-speed-race.png

@@ -0,0 +1 @@
+A clean tech dashboard illustration, 16:9 aspect ratio, showing a vertical stack of four horizontal ranking bars on a dark navy background. Each bar is labeled on the left with a model name in thin monospace type and filled to the right with a glowing cyan progress fill plus a smaller dollar-cost overlay. Top to bottom: "OPUS 4.7 — 10/10 — $0.56", "OPUS 4.6 — 10/10 — $0.44", "SONNET 4.6 — 10/10 — $0.11", "HAIKU 4.5 — 8/10 — $0.03". The top three bars are filled completely with bright electric cyan #00d4ff, glowing at the edges. The HAIKU bar is 80% filled, with the unfilled 20% rendered in dim red. A subtle scan-line grid pattern runs across the background. In the upper left corner, a small AgentHunter reticle logo (crosshair in circle with corner brackets). No human figures. Color palette: deep navy #0a1628 background, electric cyan #00d4ff bars, white monospace labels, dim red for the failure segment. Style: modern editorial tech infographic, terminal/dashboard aesthetic, high contrast, professional, minimalist. 16:9 ratio.

docs/blog/imgs/04-cost-scale.png

@@ -0,0 +1 @@
+A cinematic tech illustration, 16:9 aspect ratio, depicting two parallel horizontal "race tracks" or lanes on a dark navy blue background. The upper lane shows a brighter, more energetic streak of electric cyan light (labeled "OPUS 4.7 — 8.4s") that extends further to the right. The lower lane shows a dimmer, slightly shorter light trail (labeled "OPUS 4.6 — 9.8s"). Both trails have subtle motion-blur tails suggesting rapid movement. In the center-right, a small stopwatch icon with cyan glowing hands. A small "1.16× FASTER" badge floats near the OPUS 4.7 trail in a thin rectangular frame. The background has a subtle perspective grid vanishing into the horizon, reinforcing the sense of speed and measurement. No human figures, no vehicles — purely abstract light trails. Color palette: deep navy #0a1628 background, electric cyan #00d4ff primary (brighter for 4.7, dimmer for 4.6), white monospace labels, slight violet #6B46C1 accent. Style: modern minimalist tech infographic, abstract motion graphics, editorial quality, dark atmospheric lighting. 16:9 ratio.

docs/blog/imgs/outline.md

@@ -0,0 +1 @@
+A clean tech illustration, 16:9 aspect ratio, showing a symmetrical pan balance / weighing scale at the center of a dark navy background. On the LEFT pan sits a short tower of 3 glowing cyan stacked coin-like disks labeled "SONNET 4.6 — $0.11" with a small star/award badge floating above. On the RIGHT pan sits a much taller tower of ~15 stacked cyan coin disks labeled "OPUS 4.7 — $0.56". Both pans display a floating "10/10" green accuracy badge above their respective towers to emphasize they score identically. Between the two pans, in the upper center, floats a glowing electric cyan text label "5× CHEAPER" with a subtle arrow pointing toward the SONNET side. The balance beam tilts slightly favoring the lighter-but-equal-value SONNET side. Underneath the scale, a thin technical ruler/measurement line. No human figures. Color palette: deep navy #0a1628 background, electric cyan #00d4ff for coin stacks and highlights, emerald green #10b981 for the 10/10 accuracy badges, white monospace labels. Style: modern minimalist tech infographic, clean geometric shapes, editorial quality, symbolic and legible at a glance. 16:9 ratio.