Claude can now help build evaluations and hillclimb on them

CClaudeDevs9/28/2026, 14:00:33

Claude can now help build evaluations and hillclimb on them, ClaudeDevs says.

The same announcement shares guidance on eval design and on skills that Claude Code can use to improve applications: eval design and hillclimbing guidance.

#claude #claude-code #evaluations #eval-design

View original post ↗

Previous articleOpus 5.5 leads Anthropic models on OpenRouter by spend and tokens Next articleArena.ai ranks GPT-6 Luna (Max) 23rd in Agent Arena

Related AI news

Claude Code sets Oct 7 deadline and claim command for cloud session bonus credits

Subscribers on Claude Pro or Max plans as of Sept 23 must run /claim-credit in Claude Code by Oct 7 at 11:59pm PT to redeem their one-time cloud session credit. The credit is consumed prior to standard plan limits and expires Nov 4.

CClaudeDevs•10/6/2026, 12:04:26•

#claude-code #claude #cloud-sessions

Unused Pro, Max, and Team resets can be applied before Oct 22

ClaudeDevs says every Pro, Max, and Team user was granted a reset last Tuesday. Anyone who has not used it can still apply it at any time before Oct 22. Sonnet 5.5 is available now in the Claude Platform and Claude Code.

CClaudeDevs•9/28/2026, 11:20:46•

#claude #sonnet-5.5 #claude-code #usage-reset

Sonnet 5.5 shows up to 20% lower Claude Platform costs and more Claude Code usage

ClaudeDevs says Sonnet 5.5 has produced as much as 20% lower costs on the Claude Platform and 30% more usage in Claude Code. Claude calls it the second Claude 5.5 model, more than 30% faster than Sonnet 5 and up to 30% cheaper for most work.

CClaudeDevs•9/28/2026, 11:19:35•

#claude #sonnet-5.5 #claude-platform #claude-code

One plugin discovery experience will roll out across Claude and Claude Code

ClaudeDevs says a single discovery experience will roll out across Claude and Claude Code over the coming weeks, and describes plugins as the main way for third-party developers to extend Claude.

CClaudeDevs•9/25/2026, 13:19:38•

#claude #claude-code #plugins

ClaudeDevs says Opus 5.5 is 20% cheaper per token and 60% cheaper on cache reads

ClaudeDevs says Opus 5.5 costs 20% less per input and output token than Opus 5, and 60% less on cache reads. It also published a calculator for estimating Claude Code task cost from /usage.

CClaudeDevs•9/25/2026, 11:15:59•

#opus-5.5 #pricing #claude-code #claude

Reset history AI News FAQ Usage limits RSS feeds API MCP Skill

News is compiled from public information. The data is for reference only, and this site has no affiliation with OpenAI.

Data & original project: Did Codex Reset · Licoy