cd /news/artificial-intelligence/need-for-speed-hot-pursuit-2-ps2-dec… · home › topics › artificial-intelligence › article
[ARTICLE · art-161125] src=reversedwithai.com ↗ pub= topic=artificial-intelligence verified=true sentiment=↑ positive

Need for Speed: Hot Pursuit 2 (PS2): Decompiled Need for Speed: Hot Pursuit 2 (PS2) into C++ with Claude

A matching C++ decompilation of the PlayStation 2 version of Need for Speed: Hot Pursuit 2 (NTSC-U, SLUS-20362) has reached 9,111 of 9,230 EA-written functions, or 93.1% of the code by bytes, as of October 2026, according to the nfshp2 project's README. Anthropic's Claude models, used through Claude Code, performed the majority of the function-level decompilation, with each function accepted only after compiling with the original ee-gcc 2.9-ee-991111 compiler produced bytes matching the retail binary. The repository ships no game assets, requiring users to supply their own disc image, and a separate nfshp2-port repository builds on it for a native Apple Silicon port with widescreen and modding support.

by read1 min views1 publishedOct 30, 2026
Need for Speed: Hot Pursuit 2 (PS2): Decompiled Need for Speed: Hot Pursuit 2 (PS2) into C++ with Claude
Image: Reversedwithai (auto-discovered)

Evidence at a glance #

source evidence…

What was done #

This project provides a byte-exact C++ decompilation of the PS2 version of Need for Speed: Hot Pursuit 2 (SLUS-20362). It uses Claude models to generate code that compiles to the original retail executable using the game's original compiler.

This project is a matching decompilation of the PlayStation 2 version of Need for Speed: Hot Pursuit 2 (NTSC-U, SLUS-20362). The repository contains C++ source code designed to compile using the original game compiler, ee-gcc 2.9-ee-991111, to produce an executable that matches the retail binary byte-for-byte. The project does not include any assets or files from the original game disc; users must provide their own disc image for the build process. Anthropic's Claude models, utilized through Claude Code, performed the majority of the function-level decompilation. Each function was accepted only after the compiler produced the original bytes and passed verification checks. As of October 2026, 9,111 of 9,230 EA-written functions match, representing 93.1% of the code by bytes. The remaining code consists of Sony libraries, the C runtime, and hand-written assembly. A separate repository, nfshp2-port, uses this decompilation to provide a native port for Apple Silicon with additional features like widescreen support and modding capabilities.

Sources: Source 1

Reproduce it #

tools/setup.sh; python3 tools/extract_elf.py DISC.iso; .venv/bin/python configure.py; .venv/bin/ninja

Community #

Ask a question, share what worked, or follow the next milestone.

Your library

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @need for speed: hot pursuit 2 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/need-for-speed-hot-p…] indexed:0 read:1min 2026-10-30 · —