diff --git a/README.en.md b/README.en.md index 13758d2..313dbba 100644 --- a/README.en.md +++ b/README.en.md @@ -29,12 +29,17 @@ One-shot LLM eval cases by NAGI STUDIO: same prompt, different model x harness x | Claude Fable 5 | Anthropic | Claude Web App · Max
Claude Code · Max
Cursor · High | 06 | | Claude Opus 4.8 | Anthropic | Claude Code · Max | 02 | | Claude Opus 4.7 | Anthropic | Cursor · Max | 02 | -| Claude Opus 4.6 | Anthropic | Cursor · Max | 01 | +| Claude Opus 4.6 | Anthropic | Cursor · Max | 02 | +| Claude Opus 4.5 | Anthropic | Cursor · Thinking | 02 | | GPT-5.5 Pro | OpenAI | ChatGPT Web · Extended Pro | 02 | | GPT-5.5 | OpenAI | Codex CLI · xhigh | 02 | +| GPT-5.4 | OpenAI | Cursor · xhigh | 02 | +| GPT-5.3 Codex | OpenAI | Cursor · xhigh | 02 | +| GPT-5.2 | OpenAI | Cursor · xhigh | 02 | | Gemini 3.1 Pro | Google | AntiGravity · High
Cursor · Default
Google AI Studio · High | 04 | | Gemini 3.5 Flash | Google | AntiGravity · High
Google AI Studio · High
Cursor · Default | 04 | | Grok Build | xAI | Grok Build TUI · Max | 02 | +| Grok 4.3 | xAI | Cursor · Default | 01 | | Composer 2.5 | Cursor | Cursor · Max
Cursor · Default
Grok Build TUI · Default | 05 | | Mistral Medium 3.5 | Mistral AI | Vibe · Thinking | 02 | | DeepSeek-V4-Pro | DeepSeek | Claude Code · Max | 02 | diff --git a/README.md b/README.md index d8635c7..5f1a4bf 100644 --- a/README.md +++ b/README.md @@ -29,12 +29,17 @@ NAGI STUDIO 的 LLM 测评案例集：同一段提示词，不同「模型 × Ha | Claude Fable 5 | Anthropic | Claude Web App · Max
Claude Code · Max
Cursor · High | 06 | | Claude Opus 4.8 | Anthropic | Claude Code · Max | 02 | | Claude Opus 4.7 | Anthropic | Cursor · Max | 02 | -| Claude Opus 4.6 | Anthropic | Cursor · Max | 01 | +| Claude Opus 4.6 | Anthropic | Cursor · Max | 02 | +| Claude Opus 4.5 | Anthropic | Cursor · Thinking | 02 | | GPT-5.5 Pro | OpenAI | ChatGPT Web · Extended Pro | 02 | | GPT-5.5 | OpenAI | Codex CLI · xhigh | 02 | +| GPT-5.4 | OpenAI | Cursor · xhigh | 02 | +| GPT-5.3 Codex | OpenAI | Cursor · xhigh | 02 | +| GPT-5.2 | OpenAI | Cursor · xhigh | 02 | | Gemini 3.1 Pro | Google | AntiGravity · High
Cursor · Default
Google AI Studio · High | 04 | | Gemini 3.5 Flash | Google | AntiGravity · High
Google AI Studio · High
Cursor · Default | 04 | | Grok Build | xAI | Grok Build TUI · Max | 02 | +| Grok 4.3 | xAI | Cursor · Default | 01 | | Composer 2.5 | Cursor | Cursor · Max
Cursor · Default
Grok Build TUI · Default | 05 | | Mistral Medium 3.5 | Mistral AI | Vibe · Thinking | 02 | | DeepSeek-V4-Pro | DeepSeek | Claude Code · Max | 02 | diff --git a/models/claude-opus-4-5-cursor.json b/models/claude-opus-4-5-cursor.json new file mode 100644 index 0000000..3a4ee17 --- /dev/null +++ b/models/claude-opus-4-5-cursor.json @@ -0,0 +1,24 @@ +{ + "label": "Claude Opus 4.5", + "vendor": "Anthropic", + "harness": "Cursor", + "effort": "Thinking", + "artifactDir": "claude-opus-4-5/cursor-thinking", + "order": 18, + "runs": { + "mythos-craft": { + "note": { + "zh": "在 Cursor 中以 Claude Opus 4.5（Thinking 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by Claude Opus 4.5 (Thinking effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + }, + "pelican-cycling": { + "note": { + "zh": "在 Cursor 中以 Claude Opus 4.5（Thinking 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by Claude Opus 4.5 (Thinking effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + } + } +} diff --git a/models/claude-opus-4-6-cursor.json b/models/claude-opus-4-6-cursor.json index d42bccb..b368564 100644 --- a/models/claude-opus-4-6-cursor.json +++ b/models/claude-opus-4-6-cursor.json @@ -12,6 +12,14 @@ "en": "Generated in one shot by Claude Opus 4.6 (Max effort) inside Cursor; unmodified original output." }, "contributor": "Nagi-ovo" + }, + "pelican-cycling": { + "file": "pelican-bicycle.svg", + "note": { + "zh": "在 Cursor 中以 Claude Opus 4.6（Max 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by Claude Opus 4.6 (Max effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" } } } diff --git a/models/gpt-5-2-cursor.json b/models/gpt-5-2-cursor.json new file mode 100644 index 0000000..4c0b304 --- /dev/null +++ b/models/gpt-5-2-cursor.json @@ -0,0 +1,25 @@ +{ + "label": "GPT-5.2", + "vendor": "OpenAI", + "harness": "Cursor", + "effort": "xhigh", + "artifactDir": "gpt-5-2/cursor-xhigh", + "order": 33, + "runs": { + "mythos-craft": { + "file": "mythiccraft.html", + "note": { + "zh": "在 Cursor 中以 GPT-5.2（xhigh 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by GPT-5.2 (xhigh effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + }, + "pelican-cycling": { + "note": { + "zh": "在 Cursor 中以 GPT-5.2（xhigh 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by GPT-5.2 (xhigh effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + } + } +} diff --git a/models/gpt-5-3-codex-cursor.json b/models/gpt-5-3-codex-cursor.json new file mode 100644 index 0000000..f2e2212 --- /dev/null +++ b/models/gpt-5-3-codex-cursor.json @@ -0,0 +1,25 @@ +{ + "label": "GPT-5.3 Codex", + "vendor": "OpenAI", + "harness": "Cursor", + "effort": "xhigh", + "artifactDir": "gpt-5-3-codex/cursor-xhigh", + "order": 32, + "runs": { + "mythos-craft": { + "file": "mythic-craft.html", + "note": { + "zh": "在 Cursor 中以 GPT-5.3 Codex（xhigh 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by GPT-5.3 Codex (xhigh effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + }, + "pelican-cycling": { + "note": { + "zh": "在 Cursor 中以 GPT-5.3 Codex（xhigh 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by GPT-5.3 Codex (xhigh effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + } + } +} diff --git a/models/gpt-5-4-cursor.json b/models/gpt-5-4-cursor.json new file mode 100644 index 0000000..440ada5 --- /dev/null +++ b/models/gpt-5-4-cursor.json @@ -0,0 +1,25 @@ +{ + "label": "GPT-5.4", + "vendor": "OpenAI", + "harness": "Cursor", + "effort": "xhigh", + "artifactDir": "gpt-5-4/cursor-xhigh", + "order": 31, + "runs": { + "mythos-craft": { + "file": "index.html", + "note": { + "zh": "在 Cursor 中以 GPT-5.4（xhigh 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by GPT-5.4 (xhigh effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + }, + "pelican-cycling": { + "note": { + "zh": "在 Cursor 中以 GPT-5.4（xhigh 思考配额）一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by GPT-5.4 (xhigh effort) inside Cursor; unmodified original output." + }, + "contributor": "Nagi-ovo" + } + } +} diff --git a/models/grok-4-3-cursor.json b/models/grok-4-3-cursor.json new file mode 100644 index 0000000..92f5fb8 --- /dev/null +++ b/models/grok-4-3-cursor.json @@ -0,0 +1,17 @@ +{ + "label": "Grok 4.3", + "vendor": "xAI", + "harness": "Cursor", + "effort": "Default", + "artifactDir": "grok-4-3/cursor-default", + "order": 61, + "runs": { + "mythos-craft": { + "note": { + "zh": "在 Cursor 中以 Grok 4.3 默认设置一次生成；本页为未经修改的原始产出。", + "en": "Generated in one shot by Grok 4.3 inside Cursor with default settings; unmodified original output." + }, + "contributor": "Nagi-ovo" + } + } +} diff --git a/outputs/claude-opus-4-5/cursor-thinking/mythos-craft.html b/outputs/claude-opus-4-5/cursor-thinking/mythos-craft.html new file mode 100644 index 0000000..dd7171c --- /dev/null +++ b/outputs/claude-opus-4-5/cursor-thinking/mythos-craft.html @@ -0,0 +1,1718 @@ + + + + + + MythosCraft - A Mythical Voxel World + + + +
+

MythosCraft

+

A Mythical Voxel World of Ancient Wonders

+ +
+
WASD - Move | Space - Jump | Shift - Descend
+
Mouse - Look Around | Scroll - Select Block
+
Left Click - Break Block | Right Click - Place Block
+
1-9 - Quick Select Block
+
+
+

Generating mythical realm...

+ +
+
+

MythosCraft

+
Position: 0, 0, 0
+
FPS: 0
+
+
+ + + + + diff --git a/outputs/claude-opus-4-5/cursor-thinking/pelican-cycling.svg b/outputs/claude-opus-4-5/cursor-thinking/pelican-cycling.svg new file mode 100644 index 0000000..850c56a --- /dev/null +++ b/outputs/claude-opus-4-5/cursor-thinking/pelican-cycling.svg @@ -0,0 +1,339 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + \ No newline at end of file diff --git a/outputs/claude-opus-4-6/cursor-max/pelican-bicycle.svg b/outputs/claude-opus-4-6/cursor-max/pelican-bicycle.svg new file mode 100644 index 0000000..5b75e41 --- /dev/null +++ b/outputs/claude-opus-4-6/cursor-max/pelican-bicycle.svg @@ -0,0 +1,465 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + \ No newline at end of file diff --git a/outputs/claude-opus-4-7/cursor-max/mythos-craft.html b/outputs/claude-opus-4-7/cursor-max/mythos-craft.html index beebc1e..6ecfe8a 100644 --- a/outputs/claude-opus-4-7/cursor-max/mythos-craft.html +++ b/outputs/claude-opus-4-7/cursor-max/mythos-craft.html @@ -1,11 +1,2250 @@ - - + + - - -Mythic Realms — A Voxel Pilgrimage + + +MythosCraft · 神话之地 + -
init
+ + + +
+ +
+ +
+
+
+
+
+
+ +
+
+

MYTHOS CRAFT神 话 之 地

+

An ancient world of myths, towers, and dragon-temples

+
+
W A S D移动 / Move
+
SPACE跳跃 / Jump
+
SHIFT加速 / Sprint
+
MOUSE视角 / Look
+
L-CLICK破坏 / Break
+
R-CLICK放置 / Place
+
1 - 9选择方块 / Select
+
F飞行 / Fly
+
F3调试信息 / Debug
+
ESC暂停 / Pause
+
+ +
+ +
+ + + + + + diff --git a/outputs/gpt-5-2/cursor-xhigh/mythiccraft.html b/outputs/gpt-5-2/cursor-xhigh/mythiccraft.html new file mode 100644 index 0000000..0e7cb2c --- /dev/null +++ b/outputs/gpt-5-2/cursor-xhigh/mythiccraft.html @@ -0,0 +1,1848 @@ + + + + + + MythicCraft - Voxel Sandbox + + + +
+ + +
+
+
+
+

MythicCraft：体素沙盒（单文件）

+
+ Perlin 噪声无限分块地形 + 第一人称（指针锁）+ 拆放方块 + 内置神话幻想地形/建筑群（世界树、浮空神殿、龙骨峡谷、水晶尖塔、古代遗迹）。 +
+ MythicCraft: infinite Perlin terrain, FPS controls, voxel break/place, and built-in mythic landmarks. +
+
+
+
+ 正在准备世界… + 提示：点击后会请求鼠标锁定 / Pointer Lock +
+
+ +
+
+

操作 / Controls

+
    +
  • W A S D 移动 / Move
    • +
  • 鼠标移动 环视 / Look
    • +
  • Space 跳跃（行走模式）/ Jump (walk mode)
    • +
  • F 切换飞行 / Toggle fly
    • +
  • Shift 下落（飞行）/ Down (fly)
    • +
  • 1-6 选择方块 / Select block
    • +
  • 左键 破坏方块 / Break
    • +
  • 右键 放置方块 / Place
    • +
  • Esc 释放鼠标 / Unlock
    • +
+
+
+

神话地标 / Mythic Landmarks

+
    +
  • 世界树 / World Tree：出生点附近的巨树与符文祭坛
    • +
  • 浮空神殿 / Sky Temple：东南方向的漂浮岛屿与传送环
    • +
  • 龙骨峡谷 / Dragonbone Ravine：西北方向的骨架脊梁与肋骨拱门
    • +
  • 水晶尖塔 / Crystal Spires：西南方向的符文尖塔群
    • +
  • 古代遗迹 / Ancient Ruins：南方方向的断墙、拱门与柱廊
    • +
+
+ 你可以开启飞行快速巡游。坐标与目标方块信息会显示在左上角调试面板。 +
Use fly mode for quick sightseeing; coords and target info show in the top-left HUD.
+
+
+
+ +
+ +
+ 渲染 -- FPS + 区块 -- + 种子 -- +
+
+
+
+ + + + + diff --git a/outputs/gpt-5-2/cursor-xhigh/pelican-cycling.svg b/outputs/gpt-5-2/cursor-xhigh/pelican-cycling.svg new file mode 100644 index 0000000..cd8c811 --- /dev/null +++ b/outputs/gpt-5-2/cursor-xhigh/pelican-cycling.svg @@ -0,0 +1,390 @@ + + Pelican Cycling by the Sea + A vivid, realistic pelican riding a bicycle along a seaside beach at golden hour. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/outputs/gpt-5-3-codex/cursor-xhigh/mythic-craft.html b/outputs/gpt-5-3-codex/cursor-xhigh/mythic-craft.html new file mode 100644 index 0000000..8ab7d55 --- /dev/null +++ b/outputs/gpt-5-3-codex/cursor-xhigh/mythic-craft.html @@ -0,0 +1,1554 @@ + + + + + + Mythic Craft - Single File Minecraft Clone + + + +
+
+

Mythic Craft

+

Single-file HTML5 voxel sandbox with streaming Perlin terrain and mythic procedural landmarks.

+ + +
+
+ +
+
+
+
+ Landmarks are procedurally generated in all directions. Walk farther to discover more structures in the infinite world. +
+
+
+ + + + diff --git a/outputs/gpt-5-3-codex/cursor-xhigh/pelican-cycling.svg b/outputs/gpt-5-3-codex/cursor-xhigh/pelican-cycling.svg new file mode 100644 index 0000000..e2efa29 --- /dev/null +++ b/outputs/gpt-5-3-codex/cursor-xhigh/pelican-cycling.svg @@ -0,0 +1,276 @@ + + Pelican cycling along the seaside + A realistic and vivid pelican riding a bicycle on a coastal path at golden hour, with moving waves and soft atmospheric motion. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/outputs/gpt-5-4/cursor-xhigh/index.html b/outputs/gpt-5-4/cursor-xhigh/index.html new file mode 100644 index 0000000..04d21a7 --- /dev/null +++ b/outputs/gpt-5-4/cursor-xhigh/index.html @@ -0,0 +1,2116 @@ + + + + + + 神话方境 + + + + +
+
+
+

神话方境

+

单文件 HTML5 + Three.js 体素世界：无限分块柏林噪声地形、第一人称漫游、左键破坏、右键放置、六种可选方块与接近原版像素质感的程序化材质。

+

起始圈内置了整套神话奇境：星冠浮城、月井神殿、蚀日石门、世界树庭、坠星晶坑与龙骸山脊；继续远行时，区域还会继续生成空中圣殿、螺旋高塔、符门遗迹与晶簇天坑。

+
+
+
Realm State
+
初始化世界中...
+
正在感知地标...
+
+
WASD 移动，空格跳跃，Shift 冲刺，鼠标环视，左键破坏，右键放置，滚轮或数字键切换方块，R 返回浮城
+
+
+
+

进入神话方境

+

点击画面锁定鼠标并开始探索这个由柏林噪声与手工神话地标共同构成的体素世界。你会从星冠浮城出发，在浮空圣殿、世界树、龙骸山脊与晶石遗迹之间穿行。

+
+
+
+ + + + diff --git a/outputs/gpt-5-4/cursor-xhigh/pelican-cycling.svg b/outputs/gpt-5-4/cursor-xhigh/pelican-cycling.svg new file mode 100644 index 0000000..f925a13 --- /dev/null +++ b/outputs/gpt-5-4/cursor-xhigh/pelican-cycling.svg @@ -0,0 +1,276 @@ + + Pelican riding a bicycle by the sea + A vivid seaside illustration of a realistic pelican riding a bicycle along a glowing beach at sunset. + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/outputs/grok-4-3/cursor-default/mythos-craft.html b/outputs/grok-4-3/cursor-default/mythos-craft.html new file mode 100644 index 0000000..565c1c3 --- /dev/null +++ b/outputs/grok-4-3/cursor-default/mythos-craft.html @@ -0,0 +1,540 @@ + + + + +神话工艺 · MythosCraft + + + +
+
+
+
+ 神话工艺 MythosCraft
+ WASD 移动   空格 跳跃   鼠标 环视
+ 左键 破坏   右键 放置   1-9 选择方块
+ 点击画布 开始游戏 +
+
+
泥土 Dirt
+
草地 Grass
+
石头 Stone
+
橡木 Oak
+
叶片 Leaves
+
晶石 Crystal
+
魔能 Mana
+
+
+
+ + + + \ No newline at end of file