Leaderboard
SWE-bench Live MultiLang TypeScript/JavaScript
swebench-live-multilang-tsjsRecord
- Category
- swe
- Scored models
- 3
Scores
| Model | Score |
|---|---|
| deepseek-ai/DeepSeek-V4-Pro | 36.8 |
| deepseek-ai/DeepSeek-V4-Pro | 36.8 |
| deepseek-ai/DeepSeek-V4-Pro | 30.9 |
Remaining fields
Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.
- category
- swe
- id
- swebench-live-multilang-tsjs
- name
- SWE-bench Live MultiLang TypeScript/JavaScript
- schema version
- local-ai-registry/v1
rows
| variant | conf | context | model id | org | rank | root | score |
|---|---|---|---|---|---|---|---|
| DeepSeek-V4-Pro | high | SWE-bench Live MultiLang TypeScript/JavaScript; 75/204 resolved; OpenHands; evaluated 2026-06-05 | deepseek-v4-pro | deepseek-ai | 1 | deepseek-ai/DeepSeek-V4-Pro | 36.8 |
| DeepSeek-V4-Pro | high | SWE-bench Live MultiLang TypeScript/JavaScript; 75/204 resolved; SWE-agent; evaluated 2026-06-05 | deepseek-v4-pro | deepseek-ai | 2 | deepseek-ai/DeepSeek-V4-Pro | 36.8 |
| DeepSeek-V4-Pro | high | SWE-bench Live MultiLang TypeScript/JavaScript; 63/204 resolved; Claude Code; evaluated 2026-06-05 | deepseek-v4-pro | deepseek-ai | 3 | deepseek-ai/DeepSeek-V4-Pro | 30.9 |
Provenance & metadata (1)
source
- kind
- official-benchmark-leaderboard
- paths
- reports-0605.jsonl