Leaderboard
SWE-bench Live MultiLang Rust (2026-01-08 snapshot)
swebench-live-2026-01-08-multilang-rustRecord
- Category
- swe
- Scored models
- 3
Scores
| Model | Score |
|---|---|
| deepseek-ai/DeepSeek-V3.1-Terminus | 28.9 |
| deepseek-ai/DeepSeek-V3.1-Terminus | 26.7 |
| deepseek-ai/DeepSeek-V3.1-Terminus | 17.8 |
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-2026-01-08-multilang-rust
- name
- SWE-bench Live MultiLang Rust (2026-01-08 snapshot)
- schema version
- local-ai-registry/v1
rows
| variant | conf | context | model id | org | rank | root | score |
|---|---|---|---|---|---|---|---|
| DeepSeek-V3.1-Terminus | high | SWE-bench Live MultiLang Rust; 13/45 resolved; SWE-agent; evaluated 2026-01-08 | deepseek-v3-1-terminus | deepseek-ai | 1 | deepseek-ai/DeepSeek-V3.1-Terminus | 28.9 |
| DeepSeek-V3.1-Terminus | high | SWE-bench Live MultiLang Rust; 12/45 resolved; OpenHands; evaluated 2026-01-08 | deepseek-v3-1-terminus | deepseek-ai | 2 | deepseek-ai/DeepSeek-V3.1-Terminus | 26.7 |
| DeepSeek-V3.1-Terminus | high | SWE-bench Live MultiLang Rust; 8/45 resolved; Claude Code; evaluated 2026-01-08 | deepseek-v3-1-terminus | deepseek-ai | 3 | deepseek-ai/DeepSeek-V3.1-Terminus | 17.8 |
Provenance & metadata (1)
source
- kind
- official-benchmark-leaderboard
- paths
- reports-0605.jsonl