Inserloft commited on
Commit
5a67d67
·
verified ·
1 Parent(s): ff9fca8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +15 -11
README.md CHANGED
@@ -1,3 +1,12 @@
 
 
 
 
 
 
 
 
 
1
  # Rockman Benchmark v0.2
2
 
3
  > **A comprehensive, contamination-resistant benchmark for evaluating LLMs on code generation, debugging, optimization, and software engineering tasks.**
@@ -127,14 +136,14 @@ reporter.generate_full_report(leaderboard, detailed_results)
127
 
128
  ```json
129
  {
130
- "model_name": "qwen/qwen3.8-27b:free",
131
- "model_version": "qwen3.8-27b:free",
132
- "provider": "openrouter",
133
  "temperature": 0.0,
134
  "top_p": 1.0,
135
  "max_tokens": 4096,
136
  "k": 1,
137
- "hardware": "openrouter-cloud",
138
  "date": "2024-09-20T01:48:31Z",
139
  "evaluator_version": "0.2.0",
140
  "benchmark_version": "v0.2"
@@ -176,9 +185,6 @@ Model Pass@1 Solved
176
  Template 13.64% 6/44
177
  Heuristic 0.00% 0/44
178
  Random 0.00% 0/44
179
- qwen3.8-27b:free 4.55% 2/44 (41 rate limited)
180
- Claude Sonnet 4 — API LIMIT
181
- GPT-4o ? ?
182
  ```
183
 
184
  ---
@@ -282,7 +288,7 @@ Dataset splits:
282
  @misc{rockman2024,
283
  title={Rockman: A Comprehensive Benchmark for Code Generation and Software Engineering},
284
  author={Inserloft},
285
- year={2024},
286
  version={v0.2},
287
  url={https://github.com/InserLoft/Rockman}
288
  }
@@ -308,8 +314,6 @@ Dataset splits:
308
  | **Benchmarks** | [Inserloft.com/benchmarks/RockMan](https://Inserloft.com/benchmarks/RockMan) |
309
  | **GitHub** | [InserLoft/Rockman](https://github.com/InserLoft/Rockman) |
310
  | **Hugging Face** | [Inserloft/Rockman](https://huggingface.co/datasets/Inserloft/Rockman) |
311
- | **Email** | benchmarks@inserloft.com |
312
-
313
  ---
314
 
315
- **Built with ❤️ by [Inserloft](https://Inserloft.com)** — Advancing the science of code intelligence evaluation.
 
1
+ ---
2
+ license: mit
3
+ language:
4
+ - en
5
+ tags:
6
+ - Benchmark
7
+ - Inserloft
8
+ pretty_name: Rockman
9
+ ---
10
  # Rockman Benchmark v0.2
11
 
12
  > **A comprehensive, contamination-resistant benchmark for evaluating LLMs on code generation, debugging, optimization, and software engineering tasks.**
 
136
 
137
  ```json
138
  {
139
+ "model_name": "user/model",
140
+ "model_version": "model-version",
141
+ "provider": "provider",
142
  "temperature": 0.0,
143
  "top_p": 1.0,
144
  "max_tokens": 4096,
145
  "k": 1,
146
+ "hardware": "example-cloud",
147
  "date": "2024-09-20T01:48:31Z",
148
  "evaluator_version": "0.2.0",
149
  "benchmark_version": "v0.2"
 
185
  Template 13.64% 6/44
186
  Heuristic 0.00% 0/44
187
  Random 0.00% 0/44
 
 
 
188
  ```
189
 
190
  ---
 
288
  @misc{rockman2024,
289
  title={Rockman: A Comprehensive Benchmark for Code Generation and Software Engineering},
290
  author={Inserloft},
291
+ year={2026},
292
  version={v0.2},
293
  url={https://github.com/InserLoft/Rockman}
294
  }
 
314
  | **Benchmarks** | [Inserloft.com/benchmarks/RockMan](https://Inserloft.com/benchmarks/RockMan) |
315
  | **GitHub** | [InserLoft/Rockman](https://github.com/InserLoft/Rockman) |
316
  | **Hugging Face** | [Inserloft/Rockman](https://huggingface.co/datasets/Inserloft/Rockman) |
 
 
317
  ---
318
 
319
+ **[Inserloft](https://Inserloft.com)** — Advancing the science of code intelligence evaluation.