SWE-bench_Pro-os/index.html at main · scaleapi/SWE-bench_Pro-os · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>SWE-Bench Pro</title>
    <style>
        * {
            margin: 0;
            padding: 0;
            box-sizing: border-box;
        }

        body {
            font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif;
            background: linear-gradient(135deg, #667eea 0%, #764ba2 25%, #f093fb 50%, #4facfe 100%);
            min-height: 100vh;
            padding: 40px 20px;
        }

        .container {
            max-width: 1200px;
            margin: 0 auto;
        }

        .header {
            text-align: center;
            margin-bottom: 60px;
        }

        h1 {
            font-size: 3.5rem;
            font-weight: 800;
            color: #ffffff;
            text-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
        }

        .subtitle {
            font-size: 1.25rem;
            color: #ffffff;
            margin-top: 10px;
            text-shadow: 0 1px 2px rgba(0, 0, 0, 0.1);
        }

        .buttons {
            display: flex;
            gap: 15px;
            justify-content: center;
            margin-top: 30px;
            flex-wrap: wrap;
        }

        .btn {
            padding: 12px 28px;
            border: 2px solid #3b82f6;
            background: white;
            border-radius: 10px;
            font-size: 1rem;
            font-weight: 600;
            color: #3b82f6;
            cursor: pointer;
            transition: all 0.3s ease;
            text-decoration: none;
            display: inline-flex;
            align-items: center;
            gap: 8px;
        }

        .btn:hover {
            background: #3b82f6;
            color: white;
            transform: translateY(-2px);
            box-shadow: 0 4px 12px rgba(59, 130, 246, 0.3);
        }

        .leaderboard-section {
            background: white;
            border-radius: 20px;
            padding: 40px;
            box-shadow: 0 10px 40px rgba(0, 0, 0, 0.1);
        }

        .leaderboard-title {
            font-size: 2rem;
            font-weight: 700;
            color: #1f2937;
            margin-bottom: 30px;
        }

        table {
            width: 100%;
            border-collapse: collapse;
        }

        thead {
            background: #f3f4f6;
        }

        th {
            padding: 16px;
            text-align: left;
            font-weight: 600;
            color: #374151;
            border-bottom: 2px solid #e5e7eb;
        }

        th:nth-child(2), th:nth-child(3) {
            text-align: center;
        }

        th:last-child {
            text-align: center;
        }

        tbody tr {
            border-bottom: 1px solid #e5e7eb;
            transition: background 0.2s ease;
        }

        tbody tr:hover {
            background: #f9fafb;
        }

        td {
            padding: 16px;
            color: #1f2937;
        }

        td:nth-child(2), td:nth-child(3) {
            text-align: center;
        }

        td:last-child {
            text-align: center;
        }

        .model-name {
            font-weight: 500;
        }

        .model-icons {
            display: inline-flex;
            gap: 4px;
            margin-right: 8px;
            font-size: 1.2rem;
        }

        .resolved-score {
            font-size: 1.25rem;
            font-weight: 700;
            color: #059669;
        }

        .z-score {
            font-size: 1.1rem;
            font-weight: 600;
        }

        .z-score.positive {
            color: #059669;
        }

        .z-score.negative {
            color: #dc2626;
        }

        .link-icon {
            color: #3b82f6;
            text-decoration: none;
            font-size: 1.2rem;
            transition: transform 0.2s ease;
            display: inline-block;
        }

        .link-icon:hover {
            transform: scale(1.2);
        }

        .about-section {
            background: white;
            border-radius: 20px;
            padding: 40px;
            margin-top: 40px;
            box-shadow: 0 10px 40px rgba(0, 0, 0, 0.1);
        }

        .about-section h2 {
            font-size: 1.75rem;
            font-weight: 700;
            color: #1f2937;
            margin-top: 30px;
            margin-bottom: 15px;
        }

        .about-section h2:first-child {
            margin-top: 0;
        }

        .about-section p {
            color: #4b5563;
            line-height: 1.7;
            margin-bottom: 15px;
        }

        .about-section a {
            color: #3b82f6;
            text-decoration: none;
        }

        .about-section a:hover {
            text-decoration: underline;
        }

        .citation-box {
            background: #f9fafb;
            border: 1px solid #e5e7eb;
            border-radius: 8px;
            padding: 20px;
            margin: 20px 0;
            font-family: 'Courier New', monospace;
            font-size: 0.9rem;
            overflow-x: auto;
            white-space: pre-wrap;
            color: #1f2937;
        }

        .disclaimer {
            margin-top: 30px;
            padding-top: 20px;
            border-top: 1px solid #e5e7eb;
            color: #6b7280;
            font-size: 0.9rem;
        }

        @media (max-width: 768px) {
            h1 {
                font-size: 2.5rem;
            }

            .leaderboard-section {
                padding: 20px;
                overflow-x: auto;
            }

            .about-section {
                padding: 20px;
            }

            table {
                font-size: 0.9rem;
            }

            th, td {
                padding: 12px 8px;
            }
        }
    </style>
</head>
<body>
    <div class="container">
        <div class="header">
            <h1>SWE-Bench Pro</h1>
            <p class="subtitle">Evaluating AI Agents on Challenging Long-Horizon SWE Tasks</p>
            <div class="buttons">
                <a href="https://scale.com/leaderboard/swe_bench_pro_public" class="btn" target="_blank">🏆 Official Leaderboard</a>
                <a href="https://arxiv.org/abs/2509.16941" class="btn" target="_blank">📄 Paper</a>
                <a href="https://github.com/scaleapi/SWE-bench_Pro-os" class="btn" target="_blank">💻 Code</a>
                <a href="https://huggingface.co/datasets/ScaleAI/SWE-bench_Pro" class="btn" target="_blank">💾 Data</a>
                <a href="https://docs.google.com/forms/d/e/1FAIpQLScBjMAoPnm52nUSyOgO_iyWELuCrwjMKm0xzE2s6u-xSRBHug/viewform?usp=dialog" class="btn" target="_blank">📤 Submit</a>
            </div>
        </div>

        <div class="leaderboard-section">
            <h2 class="leaderboard-title">Leaderboard</h2>
            <table id="leaderboardTable">
                <thead>
                    <tr>
                        <th>Model</th>
                        <th>% Resolved</th>
                        <th>(+/-)</th>
                        <th>Link</th>
                    </tr>
                </thead>
                <tbody>
                    <tr data-score="43.72">
                        <td>
                            <span class="model-icons">🥇</span>
                            <span class="model-name">SWE-Agent + claude-4-5-Sonnet</span>
                        </td>
                        <td><span class="resolved-score">43.72</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                    <tr data-score="42.70">
                        <td>
                            <span class="model-icons">🥈</span>
                            <span class="model-name">SWE-Agent + claude-4-Sonnet</span>
                        </td>
                        <td><span class="resolved-score">42.70</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                    <tr data-score="39.45">
                        <td>
                            <span class="model-icons">🥉</span>
                            <span class="model-name">SWE-Agent + claude-4-5-haiku</span>
                        </td>
                        <td><span class="resolved-score">39.45</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                    <tr data-score="36.30">
                        <td>
                            <span class="model-name">SWE-Agent + gpt-5-2025-08-07 (High)</span>
                        </td>
                        <td><span class="resolved-score">36.30</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                    <tr data-score="35.52">
                        <td>
                            <span class="model-name">SWE-Agent + glm-4.5</span>
                        </td>
                        <td><span class="resolved-score">35.52</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                    <tr data-score="27.67">
                        <td>
                            <span class="model-name">SWE-Agent + kimi-k2-instruct</span>
                        </td>
                        <td><span class="resolved-score">27.67</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                    <tr data-score="16.20">
                        <td>
                            <span class="model-name">SWE-Agent + gpt-oss-120b</span>
                        </td>
                        <td><span class="resolved-score">16.20</span></td>
                        <td><span class="z-score"></span></td>
                        <td><a href="https://github.com/scaleapi/SWE-bench_Pro-os/tree/main/traj" class="link-icon" target="_blank">🔗</a></td>
                    </tr>
                </tbody>
            </table>
            <p style="margin-top: 20px; color: #6b7280; font-size: 0.95rem; line-height: 1.6;">
                Note that these results are initial runs and subject to change, pending an official announcement from Scale. Models are run with an uncapped cost and with a turn limit of 250.<br><br>
                The (+/-) column shows the 95% confidence interval (margin of error) for each score, calculated using binomial proportion statistics (total problems: 730).
            </p>
        </div>

        <div class="about-section">
            <h2>About</h2>
            <p>
                We introduce SWE-Bench Pro, a substantially more challenging benchmark that builds upon the best practices of SWE-BENCH, but is explicitly designed to capture realistic, complex, enterprise-level problems beyond the scope of SWE-BENCH. SWE-BENCH PRO contains 1,865 problems sourced from a diverse set of 41 actively maintained repositories spanning business applications, B2B services, and developer tools. The benchmark is partitioned into a public set with open access to problems sourced from 11 repositories, a held-out set of 12 repositories and a commercial set of 18 proprietary repositories where we have formal partnership agreements with early-stage startups. Problems in the held-out and the commercial set are not publicly accessible, but we release results on the commercial set. Our benchmark features long-horizon tasks that may require hours to days for a professional software engineer to complete, often involving patches across multiple files and substantial code modifications. All tasks are human-verified and augmented with sufficient context to ensure resolvability. In our evaluation of widely used coding models, under a unified scaffold, we observe that their performance on SWE-Bench PRO remains below 25% (Pass@1), with GPT-5 achieving the highest score to date at 23.3%. To better understand these limitations, we cluster the failure modes observed in the collected agent trajectories for a clearer characterization of the error patterns exhibited by current models. Overall, SWE-BENCH PRO provides a contamination-resistant testbed that more faithfully captures the complexity and diversity of real-world software development, advancing the pursuit of truly autonomous software engineering agents at a professional level.
            </p>

            <h2>Citation</h2>
            <p>
                If you found <a href="https://arxiv.org/abs/2509.16941" target="_blank">SWE-bench Pro</a> helpful for your work, please cite as follows:
            </p>

            <div class="citation-box">@inproceedings{Deng2025SWEBenchPC,
    title={SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?},
    author={Xiang Deng and Jeff Da and Edwin Pan and Yan He and Charles Ide and Kanak Garg and Niklas Lauffer and Andrew Park and Nitin Pasari and Chetan Rane and Karmini Sampath and Maya Krishnan and Srivatsa Kundurthy and Sean M. Hendryx and Zifan Wang and Chen Bo Calvin Zhang and Noah Jacobson and Bing Liu and Brad Kenstler},
    year={2025},
    url={https://api.semanticscholar.org/CorpusID:281421060}
}</div>

            <div class="citation-box">@inproceedings{jimenez2024swebench,
     title={SWE-bench: Can Language Models Resolve Real-world Github Issues?},
     author={Carlos E Jimenez and John Yang and Alexander Wettig and Shunyu Yao and Kexin Pei and Ofir Press and Karthik R Narasimhan},
     booktitle={The Twelfth International Conference on Learning Representations},
     year={2024},
     url={https://openreview.net/forum?id=VTF8yNQM66}
}</div>
            <p class="disclaimer">
            </p>
        </div>
    </div>

    <script>
        // Calculate confidence intervals for the leaderboard
        function calculateConfidenceIntervals() {
            const totalProblems = 730;
            const rows = document.querySelectorAll('#leaderboardTable tbody tr');
            const zScore = 1.96; // 95% confidence interval

            // Calculate and display confidence intervals
            rows.forEach((row) => {
                const percentage = parseFloat(row.getAttribute('data-score')) / 100;
                const solvedProblems = Math.round(percentage * totalProblems);

                // Calculate standard error for binomial proportion
                // SE = sqrt(p * (1-p) / n)
                const standardError = Math.sqrt((percentage * (1 - percentage)) / totalProblems);

                // Calculate margin of error (in percentage points)
                const marginOfError = zScore * standardError * 100;

                const ciElement = row.querySelector('.z-score');

                // Format as ± value
                ciElement.textContent = '±' + marginOfError.toFixed(2);

                // Always use positive (green) color since it's a confidence interval
                ciElement.classList.add('positive');
            });
        }

        // Run calculation when page loads
        document.addEventListener('DOMContentLoaded', calculateConfidenceIntervals);
    </script>
</body>
</html>