A business tycoon game "85% done … All built with Opus 5.5 + Sonnet 5.5"
CodeRead the post
x.com · link re-checked 2026-10-16
Claude Opus 5
By Anthropic. 14 results on 9 boards, copied as published. Different boards, different metrics: don't add them up.
License (as listed)Proprietary
Hosting regionNot confirmed
Boards9
Code
- 97%Vals AI — SWE-bench Verified
Claude Opus 5 · % resolved · board #1
Mini-SWE-agent · as of 2026-09-01 Stale ↗
- 89.03%Vals AI — LiveCodeBench
Claude Opus 5 · Accuracy % · board #4
Vals AI implementation (no agent) · as of 2026-09-01 Stale ↗
- 63.4%SWE-rebench
Opus 5 [high] · Resolved rate % · board #3
SWE-rebench harness (agent rows are full products) · as of 2026-07-01 Stale ↗
- 51.82%Terminal-Bench 4.0
Opus 5 (max) · Accuracy %
Claude Code · as of 2026-07-24 Stale ↗
- 50.3%Terminal-Bench 4.0
Opus 5 (high) · Accuracy %
Claude Code · as of 2026-07-24 Stale ↗
- 44.85%Terminal-Bench 4.0
Opus 5 (medium) · Accuracy %
Claude Code · as of 2026-07-24 Stale ↗
- 34.85%Terminal-Bench 4.0
Opus 5 (low) · Accuracy %
Claude Code · as of 2026-07-24 Stale ↗
Agentic / terminal
- 53.94%Terminal-Bench 4.0
Opus 5 (xhigh) · Resolution rate
Claude Code · as of 2026-07-24 Stale ↗
Websites
- 1691LMArena WebDev Arena
Claude Opus 5 (Max) · Arena score · board #9
LMArena WebDev (agentic front-end tasks) · as of 2026-10-22 ↗
- 1659LMArena WebDev Arena
Claude Opus 5 (High) · Arena score · board #15
LMArena WebDev (agentic front-end tasks) · as of 2026-10-22 ↗
- 88.4%Vals AI Vibe Code Bench v1.1
Claude Opus 5 · Score % · board #9
OpenHands · as of 2026-10-20 ↗
- 1306Design Arena (Fullstack)
claude-opus-5 · Elo · board #5
Design Arena · as of 2026-10-24 ↗
Design
- 1322Design Arena (All categories)
claude-opus-5 · Elo · board #13
Design Arena · as of 2026-10-24 ↗
Android
- 91.8%Android Bench 1.0 tasks (100 tasks)
Claude Opus 5 · Score %
not_stated · as of 2026-08-11 ↗
iOS
No mature public leaderboard
Available in
Use cases with this model
No sourced use case names this exact model version.
Same family: Claude 44
A Picasso-style three.js Tetris; Opus 5.5 also played and recorded the game itself
DesignRead the post
x.com · link re-checked 2026-10-16
A game "in about two weeks". GPT-6.1 Sol made the assets, Opus 5.5 did "most of the coding and game logic"
CodeRead the post
x.com · link re-checked 2026-10-16
Cost (user-reported): "look what it built through the API in 3 minutes for $0.19. Claude Opus 5.5 took 15 minutes and $2.61 on the same prompt."
DesignRead the post
x.com · link re-checked 2026-10-16
Rebuilt Shopify's four AI "gates" in headless Claude Code on tinydb and measured which gates catch bugs
CodeRead the post
dev.to · web source · link re-checked 2026-10-16
Test: "Same prompt and same references, Opus 5.5 vs Gemini 4 Argon. Opus did it in one shot. Argon got 3 rounds of feedback and still can't match…
DesignRead the post
x.com · link re-checked 2026-10-16
Sources
lmarena.ai/leaderboard/webdev vals.ai/benchmarks/vibe-code vals.ai/benchmarks/swebench vals.ai/benchmarks/lcb swe-rebench.com developer.android.com/bench tbench.ai/leaderboard designarena.ai/leaderboard designarena.ai/leaderboard/fullstack