Bug Hunt Bench
Bug Hunt Bench is A blind-graded benchmark of frontier coding models on real bugs planted in two production codebases. The score counts planted bugs only; extra fixes are tracked separately and never added.. It is ranked #705 on the Dev Radar, in AI dev tools, first seen 6 days ago and shared in 2 posts (82.8k views).
Visit bughunt.productcompass.pm
Bug Hunt Bench — frontier coding models against real planted bugs before the first paint, so a dark-theme visitor never sees a white flash. It is a file rather than an inline block because the CSP has no 'unsafe-inline'. --> Skip to the leaderboard Bug Hunt Bench Dark Blind-graded · One prompt per repo · Updated Bug Hunt Bench What this is Same prompt, one round per repo, each model in its own agentic CLI. An independent judge grades the diff against an answer key the model never sees. Method, caveats and definitions Runs on the board Featured All runs Ceiling runs only Best per lab Clear Choose runs A model can appear more than once —…
What people said about Bug Hunt Bench on X
😧惊了！实际工作实测打脸：DeepSeek V4.1 Flash 1017 仍大幅优于 Kimi K3.1 Pro！ 独立测试者 @PawelHuryn 在真实代码库上跑了 Bug Hunt 基准（2个生产级仓库 + 105个此前前沿模型漏掉的植入bug，要求找并修复）： 1️⃣DeepSeek V4.1 Flash 1017 (max)：修复 42.7/105（DeepSeek harness） 2️⃣Muse Wonder Full 1.0 (max)：42 3️⃣Opus 5.5 (max)：41.7 4️⃣Kimi K3.1 Pro (max)：仅 36/105 5️⃣Muse Spark 1.3 (max)：32.2…
— @NFT_Chen, 6 days ago · 136 likes · see the post
🤯炸了！梁文峰的实力被Kimi用“图表技巧”藏起来了！同一组标准Kimi竟然偷换数据来凸显自己！ 1️⃣Kimi官方把DeepSeek V4.1 Flash 1017画成几乎看不见的小红点，自己则高高在上。表面上比分数Kimi 65.7% vs DS 42.3%，真正被弱化的是成本，Kimi K3.1 Pro单任务成本大约是DeepSeek的20倍。 2️⃣更关键的是Kimi官方将DeepSeek的TB4.0数据给偷换了，来源不一致，表格用DeepSeek自己公布的42.3，图表却单独换了Artificial Analysis的数据。 DeepSeek开源、极便宜、迭代快； 有人选择用“小点”把它藏起来。…
— @NFT_Chen, 6 days ago · 92 likes · see the post
Alternatives to Bug Hunt Bench
- goo.gle — Officially in beta, the AppFunctions API is opening up powerful new ways to integrate your functionality.
- Kimi AI with K3.1 Pro — Try Kimi K3.1 Pro to tackle complex coding projects, build 3D models, conduct financial research, and take on more…
- Agents in flight — Give it a try
- Baseten — This session is our most in-demand yet. See you there
- project-maya — Checkout Maya
- deepinfer — You gotta love Deepseek. This is awesome stuff!
Bug Hunt Bench in numbers
- Rank on the Dev Radar: #705 of 8798
- Shared in 2 posts by 1 account: @NFT_Chen
- 82.8k views on those posts
- First seen 6 days ago, last shared 6 days ago
- Pricing seen by AGK Intelligence: not stated
- Market: AI dev tools
FAQ
What is Bug Hunt Bench?
A blind-graded benchmark of frontier coding models on real bugs planted in two production codebases. The score counts planted bugs only; extra fixes are tracked separately and never added. It was first shared on X 6 days ago and is ranked #705 on the Dev Radar.
Is Bug Hunt Bench free?
Pricing is not stated on the page we read.
Who shared Bug Hunt Bench?
1 account on X, including @NFT_Chen, in 2 posts totalling 82.8k views.
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.2k posts from 5k X accounts over the last 21 days, 8.8k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-10-30 20:27 UTC. Full method.