Qwen 3.6

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

Current reading

The readiness gap, in one scan

AI-assisted assembly · derived results

Evidence strength Strong

Decision

Human editorial judgment · 2026-06-28

Adopt with guardrails

Why: Open-weight Apache-2.0 27B at flagship-level coding scores on commodity hardware, plus an API tier that undercuts Claude Opus on cost, is a real cost/capability lever — but selective benchmark framing, unsettled Plus pricing, and tier/version sprawl all demand independent validation before trusting vendor leaderboard claims.
Next: Pilot the Apache-2.0 Qwen3.6-27B (77.2 SWE-bench Verified, ~18GB-class hardware) for self-hosted coding/tool-use against the current pipeline; A/B the Plus API on a non-PII agentic workload to confirm the 78.8 SWE-bench claim on our own evals, and pin the actual per-token price (preview vs paid) before any production swap.

Constraints

No named blocker is present in the current public projection.

Evidence summary

AI-assisted assembly

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

Publication record

Initial public reading

This is the initial public reading. No earlier readiness change is recorded.