AINewsnow

Same Haiku 5.5 Scores 82% or 65% on MCP Servers: The Harness Decides

On 1 November 2026 Sineframe published numbers from their M3 evaluation suite: the same Claude Haiku 5.5 scored 82.0% across 21 MCP servers when driven by OpenCode 1.19.5 and only 65.4% when driven by Claude Code 2.1.312. The gap was not model reasoning. It was the harness. They split the suite int…

Read the full story at DEV Community — AI ↗

Timeline · 1 report

  1. 2026-11-02 02:13 · DEV Community — AI
    Same Haiku 5.5 Scores 82% or 65% on MCP Servers: The Harness Decides

More stories

  1. Claude Code and MCP ecosystem expands with security tooling, sandboxing, and paid services — DEV Community — AI
  2. Anthropic updates Claude to stop it from recreating existing video games — Engadget
  3. Anthropic renames Claude Max plans and drops Ultra's five-hour usage limit — Anthropic (X)
  4. Anthropic hiring shifts as manager drops coding interviews and IPO questions probe candidate motives — Fortune AI
  5. Google, Microsoft, OpenAI and Anthropic to honor shared robots.txt opt-out for AI training — OpenAI News
  6. Claude Code ecosystem expands as multi-agent workflows and security concerns grow — Microsoft Research Blog
  7. Anthropic files for blockbuster IPO as Blackstone flags likely junk rating — New York Times Technology
  8. Anthropic research explores automating detection of subtle emergent misalignment — Anthropic (X)

Get the daily brief of stories like this at 6:30 every morning →