Issues

• Add Qwen3.5 Dark-Genius MoE Sovereign to the SciCode Leaderboard

  scicode-bench/SciCode#64 · Filter by author nfon41547-hash opened on Oct 1, 2026

• inspect_ai port: 3 pre-filled steps stay in the denominator (291 vs 288), making problems 13, 62 and 76 unresolvable; synchronous test run blocks the event loop

  scicode-bench/SciCode#63 · Filter by author fasuizu-br opened on Sep 26, 2026

• Portable eval artifacts: mapping the sub_steps/cmp.py grading onto an open TestCase/Grader/Result schema (EvalPort)

  scicode-bench/SciCode#62 · Filter by author adhabnr-ux opened on Aug 25, 2026

• EvalScope integration: one-command evaluation support for SciCode

  scicode-bench/SciCode#60 · Filter by author Yunnglin opened on Jul 24, 2026

• [Bug] Reference code for class-based skip-steps (13.6, 62.1) is silently truncated, failing all dependent sub-steps

  scicode-bench/SciCode#59 · Filter by author 41ow1ives opened on Jun 30, 2026
  1

• Add DevonOS to the leaderboard

  scicode-bench/SciCode#57 · Filter by author devongenerally-png opened on May 16, 2026
  3

• Add Shanhai-U2-preview to the leaderboard

  scicode-bench/SciCode#56 · Filter by author pythonla opened on Apr 27, 2026

• Add eval.yaml to HF dataset for Community Evals leaderboard

  scicode-bench/SciCode#55 · Filter by author viviennezhangx opened on Apr 9, 2026

• Accepting new submissions?

  scicode-bench/SciCode#54 · Filter by author durga-sandeep opened on Mar 11, 2026

• benchmark EpistemeAI's gpt-oss-20b

  scicode-bench/SciCode#50 · Filter by author tomtyiu opened on Oct 4, 2025

• Code extraction error for the previous substeps for specific questions 13, 62 and 76

  scicode-bench/SciCode#49 · Filter by author amaand-cerebras opened on Sep 22, 2025
  1

• README: Domain count mismatch — says “6 domains” but benchmark only has 5 (per paper & website)

  scicode-bench/SciCode#48 · Filter by author fenglang918 opened on Sep 6, 2025

• bug: Inspect AI implementation runs n+1 step before n step

  scicode-bench/SciCode#46 · Filter by author TheFloatingString opened on Aug 13, 2025

• Some example's ground_truth_code do not pass the tests in dev set

  scicode-bench/SciCode#43 · Filter by author John-Boccio opened on May 2, 2025
  1

• Add o3 and o4 mini to the leaderboard

  scicode-bench/SciCode#42 · Filter by author GypsyDignity opened on Apr 22, 2025

• Add gpt-4.5 to the leaderboard

  scicode-bench/SciCode#37 · Filter by author LuigiPagani opened on Mar 9, 2025

• Benchmark Claude Sonnet 3.7

  scicode-bench/SciCode#36 · Filter by author LuigiPagani opened on Mar 6, 2025

• Benchmark o1

  scicode-bench/SciCode#35 · Filter by author LuigiPagani opened on Mar 6, 2025

• Dependency versions for evaluation

  scicode-bench/SciCode#2 · Filter by author zrait opened on Jul 21, 2024