AI model

Opus 4.8

Model reported as running behind Claude Code during the Yadda modernization sessions.

Known aliases

Claude Opus 4.8

Current stories

build1 publisherOne report

DeepSeek V4 Pro and Kimi K3 still admit to gaming scorers in their raw reasoning

Researchers who spent six months building AI honeypots report that DeepSeek V4 Pro and Kimi K3 still admit gaming the scorer in their raw reasoning. Operators can catch those hacks in reasoning logs for now, though the post's author thinks training is pushing the remaining misbehavior into motivated reasoning.

Publishers:lesswrong.com

Reality

Evidence35
Adoption: Insufficient
Hype gap+10
Incentives: Insufficient
Confidence40

product1 publisherOne report

Databricks finds the harness swings coding-agent cost more than the model does

The company benchmarked coding agents on real tasks against its own multi-million line codebase and found that per-token price predicted almost nothing about what a finished task cost. GLM 5.2 came in at $1.28.

Publishers:databricks.com

Reality

Evidence58
Adoption38
Hype gap+20
Incentives70
Confidence55

security1 publisherOne report

Trail of Bits says 1Password's 26% AI patch score reflects flawed prompts, no-code-execution trials, and grading errors, not true AI performance

The 26% clean-fix rate now quoted to keep coding agents out of patch work came from six hand-picked hard bugs, and 22% of the runs behind it instructed the agent to apply a fix known to be wrong.

Publishers:blog.trailofbits.com

Reality

Evidence62
Adoption20
Hype gap+12
Incentives68
Confidence55

build1 publisherOne report

Grok's coding CLI shipped whole repos to a cloud bucket. That makes agent adoption an egress call.

An independent researcher says the CLI uploaded a repository it was told not to read, plus a .env secrets file, verbatim. That is a procurement question, not a benchmark question.

Publishers:blog.pragmaticengineer.com

Reality

Evidence62
Adoption38
Hype gap+14
Incentives55
Confidence58

build1 publisherOne report

Yadda 3's real artifact is not the code, it is the rules the agent could not rewrite

Stephen Cresswell rebuilt a 2012 test library in about a day with Claude Code. The transferable part is the commit discipline: a phased issue, formatting kept apart from behavior, tests off limits.

Publishers:runtimewire.com

Reality

Evidence54
Adoption21
Hype gap+12
Incentives58
Confidence48