Search-SWE
GPU TASK Valid result

Khmer-to-English Embedding Training

DeepSeek V4.1 Flash 1017 · Codex (v0.164.0) · reasoning effort max

Task description ↗Task at evaluation revision ↗← Previous taskNext task →
Task score (%)20.9
Input Tokens55,183,710including 54,538,752 cached
Output Tokens230,635
Agent time5h 55min
Cost$0.40
Recorded steps334

Cost is estimated from recorded agent tokens at official API rates in USD (2026-11-01; DeepSeek off-peak pricing).

Loading trajectory…

Select this tab to browse the saved files.

Coding time & token usage
Original agent completion
normal
Coding started
2026-10-24T14:23:13.019014Z
Coding finished
2026-10-24T20:18:43.121994Z
Input (includes cache)
55,183,710
Cached input (subset)
54,538,752
Cache writes (subset)
0
Output
230,635
Reasoning (reported separately; not added)
101,911
Recorded responses
331

Input includes cached input; output is counted once. Auxiliary reasoning counts are shown as recorded. Regrading does not repeat coding usage.

Select this tab to browse the saved files.

Configuration

Evaluation settings summary
{
  "agent": "Codex",
  "agent_timeout_sec": 21600.0,
  "coding_commit": null,
  "effort": "max",
  "model": "deepseek/deepseek-v4.1-flash-1017",
  "provider": "OpenRouter",
  "resources": {
    "cpus": 8,
    "gpus": 1,
    "memory_mb": 32768,
    "storage_mb": 102400
  },
  "run_mode": "official_harbor_full",
  "task_commit": "bb34e915be5c2b2948146c2d7862cb58652d6300",
  "verifier_commit": "bc2ef55cf5bc4e5133c32816a7874368192a2f1e",
  "version": "0.164.0"
}

Select this tab to browse the saved files.

Resolved configuration

Select this tab to browse the saved files.

Trial log

Select this tab to browse the saved files.

Exceptions

Select this tab to browse the saved files.

Select this tab to browse the saved files.

Trajectory downloads & provenance
Exported ATIF trajectory (.json.gz) ↓Final evaluation (.json) ↓

Public text redacts private values and normalizes host paths. Source hashes identify the original records.

Trial ID
khmer-embedding-training-deepseek-v4.1-flash-1017-max-07944be61674
Source trajectory SHA-256
12eb7e8d5d37098e845239e73ac1920c0d869c88d5ff44f106dc9a08a62a3b46
Final reward SHA-256
ae0a493980c1c4ec5a95b08cd11e69ee5b022b65307417757a4e458e436d6ce3