Sonnet vs Opus: Which AI Thinks Like a Content Strategist?
By Lisa Peyton and her team of AI Super Heroes
A Comparative Analysis for Content Marketers
Models tested: Claude Sonnet 4.5 | Claude Sonnet 4.6 | Claude Opus 4.6 Task: Generate infographic concepts for a B2B edtech client presenting at an industry conference
The Quick Scorecard
| Criteria | Sonnet 4.5 | Sonnet 4.6 | Opus 4.6 |
|---|---|---|---|
| Concepts generated | 6 | 9 | 10 |
| Strategic depth | ⭐⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐⭐ |
| Client context awareness | ⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐⭐ |
| Production readiness | ⭐⭐⭐ | ⭐⭐⭐⭐⭐ | ⭐⭐⭐⭐ |
| Creative range | ⭐⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐ |
| Lead gen thinking | ⭐⭐ | ⭐⭐⭐⭐⭐ | ⭐⭐⭐⭐ |
| Confidence of recommendations | ⭐⭐ | ⭐⭐⭐⭐ | ⭐⭐⭐⭐⭐ |
Sonnet 4.5 — The Competent Generalist
Sonnet 4.5 delivered a solid, well-organized portfolio of 6 concepts with clear headlines, visual approaches, and data points. The structure is clean and the concepts are logically sound. For a first pass from a junior team member, this would be a perfectly acceptable draft.
Where it falls short is strategic altitude. The output reads like a model that understood the category of the task but not the specific client situation. The recommendations feel template-driven — “bridge metaphor,” “before/after,” “myth-busting” — rather than derived from actual insight into the client’s positioning challenge. There’s also a telling tell at the end: it closes by asking three questions about what to do next. That’s a junior move. A strategic partner comes with a point of view.
Bottom line for content marketers: Good enough to spark ideas. Not good enough to hand to a client.
Sonnet 4.6 — The Production-Ready Strategist
This is a meaningful leap. Sonnet 4.6 generated 9 concepts with noticeably sharper strategic framing and — critically — production awareness that the other two models lack. It’s the only output that references the actual output destination, includes specific CTA copy and placement guidance, and frames each concept with a “Why It Works” rationale that speaks to audience psychology, not just visual design.
The lead-gen thinking here is particularly strong. One concept was explicitly framed as a conversion tool — a checklist mechanic designed to trigger self-identification and a natural demo request. That’s content strategy, not just content ideation. Sonnet 4.6 also correctly identifies which concepts have digital repurposing legs beyond the conference, which shows it’s thinking about content ROI across channels.
The prioritization guidance at the end is confident and actionable: it names a “must-have anchor piece” without hedging.
Bottom line for content marketers: This is the model you use when you need to move fast from brief to deliverable. It thinks in terms of client workflow, not just content quality.
Opus 4.6 — The Strategic Advisor in the Room
Opus 4.6 is doing something qualitatively different from the other two, and it shows up immediately in the opening paragraph. Before generating a single concept, it flags a critical production constraint — the AI image gen tool in use meant text overlays needed to be handled in post — and adjusts its creative recommendations accordingly. Neither other model noticed or addressed this at all. That’s the difference between a tool and a thinking partner.
The 10 concepts Opus generates are slightly less detailed in layout specs than Sonnet 4.6, but they’re more contextually intelligent. Opus is the only model that attributes specific concepts to things the client stakeholder actually said in the meeting, referencing her stated priorities and connecting concepts back to the core educational argument she needed to make. It’s also the only model that references the existing sales deck and flags that one concept mirrors brand assets already in use — valuable context that prevents redundant work.
The pairing recommendations at the end reveal the sharpest strategic instinct of the three: Lead with the problem, educate with the mechanism, close with proof and CTA. That’s a classic conversion narrative arc, and Opus names it intuitively without being asked to.
Where Opus is slightly weaker than Sonnet 4.6 is in granular production detail — CTA copy, placement specs, repurposing notes. It thinks at the campaign level; Sonnet 4.6 thinks at the asset level.
Bottom line for content marketers: This is the model you use for client-facing strategic concepting where context retention and judgment matter more than execution speed.
The Bigger Takeaway for Your Workflow
These outputs reveal something important about how to deploy these models strategically rather than interchangeably.
Sonnet 4.5 is your brainstorm accelerator for early-stage ideation when you need volume and a starting point quickly.
Sonnet 4.6 is your production partner — use it when you’re moving from strategy to execution and need output that’s close to client-ready without heavy editing.
Opus 4.6 is your strategic advisor — use it when you have rich context to feed it (meeting notes, brand docs, prior conversations) and need it to synthesize that context into recommendations with real judgment behind them. It rewards detailed briefing more than the others do.
The most sophisticated workflow? Use Opus to set strategy and prioritize, then hand the winning concepts to Sonnet 4.6 to build out the production-ready specs. That’s not model competition — that’s an actual AI content team.
Prompt quality reminder: All three models were working from the same brief, but Opus 4.6’s output suggests it is simply better at extracting and applying nuance from complex inputs. The lesson for content marketers: the model ceiling is only as high as your brief allows.
