Codex Results by Model
| # | Model | Run | Agent | Mode | Score | Eval | Tests | Est. cost | Calls | Wall | Evidence |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 58.7% | ok | 262/446 | $1.68 | 38 | 0.11h | manifest · eval json · eval summary · agent summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 57.0% | ok | 254/446 | $3.20 | 56 | 0.15h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 46.2% | ok | 206/446 | $2.85 | 61 | 0.15h | not exported |
Why Scores Differ
Official ProgramBench rows are the public mini-SWE-agent submissions. GoalBench rows are separate Codex /goal submissions. Same model label does not mean the same agent, prompt, tool loop, or generated implementation. A GoalBench row is resolved only when the ProgramBench-scored pass rate is exactly 100%.
GPT 5.5 (xhigh) · 20260518T232006Z
58.7% from 262/446 ProgramBench-scored tests. Raw public eval statuses: error: 1, failure: 229, passed: 298, skipped: 8.
Agent trace: 82 shell command(s), 61 executable-observation command(s), 3 build command(s), 2 blocked attempt(s). Raw Codex logs are not published.
Likely miss class: behavioral mismatch, exact version/output mismatch, terminal rendering mismatch.
06b3cee11410tests.test_config.test_version_output_format (failure)06b3cee11410tests.test_tui_advanced.test_tui_very_long_lines (failure)06b3cee11410tests.test_tui_comprehensive.test_tui_search_execution (failure)06b3cee11410tests.test_tui_comprehensive.test_tui_navigation_down_up (failure)06b3cee11410tests.test_tui_comprehensive.test_tui_home_end_keys (failure)0c9541bb06catests.test_config.test_config_precedence_full_chain (failure)
manifest · eval json · eval summary · agent summary · usage audit
GPT 5.5 (xhigh) · 20260517T094700Z
57.0% from 254/446 ProgramBench-scored tests. Raw public eval statuses: error: 1, failure: 237, passed: 290, skipped: 8.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch, interactive terminal behavior mismatch, terminal rendering mismatch.
06b3cee11410tests.test_config.test_version_output_format (failure)0c9541bb06catests.test_config.test_config_precedence_full_chain (failure)0c9541bb06catests.test_cli.test_help_equals_syntax_rejected (failure)0c9541bb06catests.test_cli.test_empty_string_project_root (failure)0c9541bb06catests.test_cli.test_combined_short_flags_help_wins (failure)0c9541bb06catests.test_config.test_default_config_when_no_file_exists (failure)
GPT 5.5 (xhigh) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
Official ProgramBench Results by Model
| # | Model | Provider | Score | Cost | Calls |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.7 (xhigh) | Anthropic | 65.0% | $2.51 | 79 |
| 2 | GPT 5.5 (xhigh) | OpenAI | 62.1% | $6.46 | 83 |
| 3 | GPT 5.5 (high) | OpenAI | 61.9% | $4.18 | 61 |
| 4 | GPT 5.4 | OpenAI | 61.0% | $0.36 | 13 |
| 5 | Claude Opus 4.7 | Anthropic | 60.8% | $2.92 | 96 |
| 6 | GPT 5.5 | OpenAI | 60.3% | $0.91 | 22 |
| 7 | Claude Haiku 4.5 | Anthropic | 54.3% | $0.64 | 95 |
| 8 | Gemini 3.1 Pro | 52.9% | $5.31 | 134 | |
| 9 | Gemini 3 Flash | 48.4% | $0.44 | 116 | |
| 10 | GPT 5 mini | OpenAI | 41.3% | $0.01 | 10 |
| 11 | GPT 5.4 mini | OpenAI | 41.0% | $0.03 | 15 |
| 12 | Claude Opus 4.6 | Anthropic | 40.8% | $12.94 | 210 |
| 13 | Claude Sonnet 4.6 | Anthropic | 6.5% | $17.28 | 310 |