Live Dashboard of PRs collections and SWE Tasks Generation
Live state of GitHub PR collection and verifiable SWE-Bench task generation across 8 languages. Detailed per-language analysis is collapsed below — click any section to expand.
Language Progress
| Language | PRs collected | Last 1h | Last 24h | Valid SWE | Last 1h | Last 24h | Processed | Success rate |
|---|---|---|---|---|---|---|---|---|
| Cc | 29,471 | -2,886 | -2,886 | 10,271 | +562 | +562 | 29,171 | |
| C++cpp | 45,625 | -2,363 | -2,363 | 4,062 | +23 | +23 | 45,552 | |
| Gogo | 126,449 | -5,654 | -5,654 | 8,473 | +378 | +378 | 125,524 | |
| Javajava | 84,869 | -4,160 | -4,160 | 4,372 | +355 | +355 | 84,386 | |
| JavaScriptjs | 36,409 | -3,273 | -3,273 | 7,257 | +120 | +120 | 36,109 | |
| Pythonpy | 98,883 | -6,885 | -6,885 | 5,971 | +913 | +913 | 98,321 | |
| Rustrust | 68,650 | -3,215 | -3,215 | 5,595 | +90 | +90 | 68,018 | |
| TypeScriptts | 66,733 | -3,757 | -3,757 | 6,846 | +441 | +441 | 66,072 |
Run Parameters
| Language | Eval model (OPENAI) | Completion model (ANTHROPIC) | Concurrency | min_source_files | max_source_files |
|---|---|---|---|---|---|
| C | Qwen3.6-35B-A3B | claude-opus-4-8 | 12 | 2 | 15 |
| C++ | Qwen3.6-35B-A3B | claude-opus-4-8 | 8 | 2 | 15 |
| Go | Qwen3.6-35B-A3B | claude-opus-4-8 | 12 | 2 | 10 |
| Java | Qwen3.6-35B-A3B | claude-opus-4-8 | 8 | 2 | 10 |
| JavaScript | Qwen3.6-35B-A3B | claude-opus-4-8 | 12 | 2 | 10 |
| Python | Qwen3.6-35B-A3B | claude-opus-4-8 | 12 | 3 | 15 |
| Rust | Qwen3.6-35B-A3B | claude-opus-4-8 | 8 | 2 | 10 |
| TypeScript | Qwen3.6-35B-A3B | claude-opus-4-8 | 12 | 2 | 10 |
Failure Reason Breakdown
click to expand
| Language | Processed | Valid SWE | Failed | trivial_pr | validation | infra_error | timeout | workflow_error | Other |
|---|---|---|---|---|---|---|---|---|---|
| C | 29,171 | 10,271 | 18,900 | 16,541 | 2,748 | 259 | 1 | 127 | 2 |
| C++ | 45,552 | 4,062 | 41,490 | 8,660 | 5,316 | 28,905 | 159 | 452 | 266 |
| Go | 125,524 | 8,473 | 117,051 | 39,618 | 25,849 | 49,173 | 1,393 | 1,006 | 18 |
| Java | 84,386 | 4,372 | 80,014 | 25,867 | 10,835 | 37,672 | 1,155 | 3,364 | 1,599 |
| JavaScript | 36,109 | 7,257 | 28,852 | 19,465 | 1,535 | 9,392 | 1 | 103 | 0 |
| Python | 98,321 | 5,971 | 92,350 | 43,623 | 12,359 | 36,744 | 857 | 469 | 120 |
| Rust | 68,018 | 5,595 | 62,423 | 29,531 | 2,480 | 28,109 | 1,071 | 665 | 1,233 |
| TypeScript | 66,072 | 6,846 | 59,226 | 19,981 | 11,751 | 25,551 | 1,776 | 541 | 8 |
trivial_pr: the PR was judged by the LLM as too trivial (e.g. only config, docs, or dependency-version changes) and unsuitable as a SWE task.
validation: validation failed after task generation (the NOP agent did not return reward=0, or the ORACLE agent did not return reward=1).
infra_error: infrastructure error (Docker build failure, network timeout, insufficient disk space, etc.).
timeout: processing timed out (per-PR total timeout or Claude Code session timeout).
workflow_error: workflow error (PR metadata fetch failure, worktree creation failure, patch generation failure, etc.).
fix.patch Complexity
click to expand
| Language | Valid SWE Count | Avg fix.patch lines | Avg fix.patch hunks | Avg fix.patch files |
|---|---|---|---|---|
| C | 10,271 | 333.26 | 17.96 | 5.82 |
| C++ | 4,062 | 285.93 | 13.74 | 5.09 |
| Go | 8,473 | 210.55 | 12.49 | 4.32 |
| Java | 4,372 | 166.81 | 10.83 | 4.31 |
| JavaScript | 7,257 | 77.11 | 6.31 | 2.80 |
| Python | 5,971 | 156.52 | 11.23 | 3.95 |
| Rust | 5,595 | 226.35 | 13.19 | 4.11 |
| TypeScript | 6,846 | 154.60 | 9.52 | 4.11 |
Metric Definitions
click to expand
Difficulty score (difficulty_score)
Reads each valid task directory's solution/fix.patch, tests/, and instruction.md, scored statically with zero API calls by src/swegen/scoring.py.
The current formula uses log-scale continuous scoring to avoid mid-sized patches becoming hard too early. Weights: patch_scope 38%, logic_complexity 32%, context_breadth 15%, test_complexity 10%, instruction_complexity 5%.
Label thresholds: easy <= 4.0, medium <= 7.0, hard > 7.0.
Tag generation and display
tags are not computed live by the dashboard; they are generated by the LLM from PR information when swegen builds the task, and written to [metadata].tags in task.toml.
The prompt asks for tags in four parts: programming language, project layer/domain, framework/library or specific topic, and a domain-independent bug class (e.g. missing-fallback, incomplete-validation). The dashboard reads existing task.toml files, counts each language's tag occurrences and share, and treats the 4th tag as the bug class for the Bug-Class panels below.
fix.patch statistics
Patch stats come from each valid task's solution/fix.patch, filtering code files by language extension, consistent with the code-only stats in upload_march_swe_to_hf.py.
Avg fix.patch lines counts added/removed lines in code-file diffs; Avg fix.patch hunks counts @@ hunks; Avg fix.patch files counts the code files involved.
difficulty_label Distribution
click to expand
| Language | easy / medium / hard | easy | medium | hard |
|---|---|---|---|---|
| C | 897 / 6768 / 2598 | 897 | 6,768 | 2,598 |
| C++ | 433 / 2524 / 1100 | 433 | 2,524 | 1,100 |
| Go | 705 / 6101 / 1661 | 705 | 6,101 | 1,661 |
| Java | 472 / 2824 / 1072 | 472 | 2,824 | 1,072 |
| JavaScript | 1096 / 5376 / 784 | 1,096 | 5,376 | 784 |
| Python | 281 / 3764 / 1902 | 281 | 3,764 | 1,902 |
| Rust | 391 / 3312 / 1890 | 391 | 3,312 | 1,890 |
| TypeScript | 627 / 5056 / 1163 | 627 | 5,056 | 1,163 |
difficulty_score Overview
click to expand
| Language | count | min | p25 | median | mean | p75 | max |
|---|---|---|---|---|---|---|---|
| C | 10,263 | 2.4 | 4.9 | 6.0 | 5.99 | 7.1 | 9.2 |
| C++ | 4,057 | 2.5 | 4.9 | 6.0 | 5.99 | 7.2 | 9.1 |
| Go | 8,467 | 2.6 | 4.9 | 5.8 | 5.84 | 6.8 | 9.1 |
| Java | 4,368 | 2.8 | 4.8 | 5.9 | 5.91 | 7.0 | 9.2 |
| JavaScript | 7,256 | 2.6 | 4.4 | 5.2 | 5.36 | 6.2 | 9.2 |
| Python | 5,947 | 2.6 | 5.3 | 6.3 | 6.29 | 7.3 | 9.1 |
| Rust | 5,593 | 2.7 | 5.2 | 6.3 | 6.26 | 7.4 | 9.0 |
| TypeScript | 6,846 | 2.7 | 4.7 | 5.6 | 5.72 | 6.6 | 9.2 |
Global Top Tags
click to expand
Per-Language Tag Distribution
click to expand
Global Top Bug Classes
click to expand
Bug class is the 4th tag in task.toml -> [metadata].tags: a domain-independent label describing the defect mechanism (e.g. missing-fallback, incomplete-validation, off-by-one-error). Generated by the LLM during swegen create and backfilled into legacy 3-tag tasks via swegen backfill-tags.
Per-Language Bug-Class Distribution
click to expand
Top bug classes per mainstream language. Counts are over tasks whose task.toml already carries a 4-tag entry; tasks still on the legacy 3-tag schema do not contribute until the backfill catches up.