pragma.vision Technology observatory

Verification register AI & Agents

Readiness verdict

Claude 3 family + Claude 3.5 Sonnet (Anthropic)

A dated reading of what is claimed, reported, and independently verified in the current evidence.

As of
2026-07-23
Revision
1
Method
v1.3.0

Current reading

Readiness band and full integer triple

AI-assisted assembly · derived results

Readiness band

Unknown

Primary summary from verified readiness

Confidence
54% · stale
Computed at
2026-07-23T19:13:04.856178+00:00
Claimed
10

Public ambition and stated capability

Reported
8

Observed practitioner reporting

Verified
5

Independently supported evidence

Gap
+5

Claimed leads verified

Evidence strength Weak

Decision

What the current evidence supports

Human editorial judgment · 2026-07-23

Track; not yet

Why
Every model in this generation (Opus/Sonnet/Haiku 3, Sonnet 3.5 v1/v2, Haiku 3.5) is now Retired on the Claude API per Anthropic's official deprecation table — the generation is fully superseded and non-functional in production, so this is a historical lineage entry, not a live adoption decision.
Next
No adoption action needed — the ecosystem already runs current Claude generations (Opus 4.8, Sonnet 5). Keep this record for register/lineage completeness only; if any surface still hardcodes a claude-3-* or claude-3-5-* model ID, migrate it immediately since requests to retired models fail outright.

Constraints

Blockers

No named blocker is present in the current public projection.

Evidence summary

Derived counts

AI-assisted assembly

Total
4
Tier 1
0
Tier 2
0
Tier 3
4
Supports
2
Contradicts
1
Context
1
Latest observed
2026-04-20

Counts and dates only. Raw signals, private excerpts, trust records, and internal corpus material are not published here.

Publication record

Revisions

Initial public reading

  1. 2026-07-23 Reading moved from unknown to unknown.

Your opinion

Tell us anything.

What works, what doesn't, what's missing — especially about our watches, lenses, and the register itself. Anonymous is fine; leave an email if you'd like a reply.