OverviewCompare › qwen3-coder:30b vs GPT-5.5 (Codex)
Head-to-head

qwen3-coder:30b vs GPT-5.5 (Codex)

Both models answered the same one-shot tasks under identical rules — no retries, answers published unedited. Two separate axes, never blended: does the code actually run, and how do blind-scored text answers hold up against a written gold answer. Frontier references sit outside the blind field and are deliberately unscored on text.

ModelRunnable simsBlind text avgFailure modes
qwen3-coder:30b 25% (21/85)9.34 (38 scored)ERR 36 · STAT 9 · BLANK 18 · HANG 1
GPT-5.5 (Codex) frontier reference 95% (81/85)unscored by designERR 2 · STAT 1 · BLANK 1

Biggest contrasts

Rounds where the two models diverged most — open them and read both answers side by side.

Roundqwen3-coder:30bGPT-5.5 (Codex)
R155 · Linked list with animated pointer relinkingSTATOK
R152 · Hash table with visible collision chainsERROK
R151 · Binary search tree growing node by nodeERROK
R139 · Small n-body system orbiting a central starOKERR
R137 · Cloth mesh hanging and rippling in the windSTATOK

→ Interactive round-by-round comparison · → Full leaderboard

Generated 2026-07-31 from the published rounds. Harness failures are excluded from every base — a broken cluster node is our fault, not the model's.