Anthropic scored 4.2/10.

Anthropic’s strongest dimension is Post-click.

Post-click scores 6.4 / 10. The dim covers 4 signals in the rubric; the page still has 1 finding in this area, but the overall score is strong relative to peers.

Anthropic vs the rest of foundation model.

1. 01Replicate6.0/10
2. 02Hugging Face5.6/10
3. 03AnthropicYou are here4.2/10

What’s costing Anthropic, quoted from the page.

1. 01
   Single-model showcase conflicts with buyers wanting multi-LLM access

   “Try Claude / Claude Opus 4.7 / Introducing a smarter, more capable Opus for coding, agents, vision, and complex professional work.”

   The page exclusively highlights Claude (Claude Opus 4.7, Claude on Mars, 'Try Claude' CTA), while 72% of buyers explicitly ask to 'access a wide variety of large language models (LLMs) in one place' — a need the page neither addresses nor positions against.

2. 02
   The page has no proof from enterprise buyers, developers, or institutions — the exact audience the rest of the page signals.

   “No customer logos, no named enterprise deployments, no developer testimonials, no G2 or analyst citations anywhere on the page. The strongest real-world proof (Claude on Mars) is formatted as a dated announcement card, not a hero-level trust signal. Three separate modules — aware…”

   The page has no proof from enterprise buyers, developers, or institutions — the exact audience the rest of the page signals. The NASA Mars story exists but reads as a press release, not a trust signal. Visitors who arrive to evaluate Anthropic for serious deployment leave without a single named customer, case study, or third-party validation.

3. 03
   The page never names what Anthropic's safety positioning actually means for a buyer's risk.

   “The headline states the safety position. The subheadline restates it in mission language. The body links to policy documents and research papers. Nowhere does the page say: 'Here is what this means for your deployment, your compliance posture, or your liability exposure.' Competi…”

   The page never names what Anthropic's safety positioning actually means for a buyer's risk. 'Safety at the frontier' is a claim; it doesn't tell an enterprise procurement team, a government agency, or a developer what they get that they wouldn't get from OpenAI. The differentiation exists in Anthropic's research — Constitutional AI, the Responsible Scaling Policy, Claude's Constitution — but none of it is translated into buyer-facing outcomes.

4. 04
   The headline names what Anthropic is, not who Claude is for.

   “AI research and products that put safety at the frontier | AI will have a vast impact on the world. Anthropic is a public benefit corporation dedicated to securing its benefits and mitigating its risks.”

   The headline describes the company's mandate, not the buyer's outcome. "Research and products" covers everyone from policy academics to API developers, naming no one. For product-aware builders comparing Claude to GPT-4, the above-fold gives no signal about what Claude actually does for their work.

5. 05
   Your subheadline opens by telling buyers what they already know.

   “AI will have a vast impact on the world.”

   Your subheadline opens with a premise every single visitor already accepts. That sentence spends the highest-attention line on the page confirming a belief instead of advancing one — and it delays the only claim that actually separates Anthropic from OpenAI or Google: what safety-first AI research produces for the people building with it.