849generated behavioral tests
80.9%official best score
69.8%Codex best score
3Codex result rows
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 | 69.8% | ok | 593/849 | n/a | 59 | 0.15h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 66.1% | ok | 561/849 | $5.68 | 71 | 0.31h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 65.1% | ok | 553/849 | n/a | 48 | 0.17h | 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
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
66.1% from 561/849 ProgramBench-scored tests. Raw public eval statuses: failure: 301, passed: 669, skipped: 3.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
642420bd4df9tests.test_allow_deny.test_deny_nonexistent_file (failure)642420bd4df9tests.test_config_gaps.test_config_missing_cache_dir_causes_error (failure)642420bd4df9tests.test_dump_check_required.test_check_required_parent_traversal_rejected (failure)642420bd4df9tests.test_cmd_edge_cases.test_allow_migration_from_old_location (failure)642420bd4df9tests.test_dotenv.test_unclosed_quote_error (failure)642420bd4df9tests.test_dump_check_required.test_dump_zsh_format_syntax (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 | GPT 5.5 (xhigh) | OpenAI | 80.9% | $6.28 | 45 |
| 2 | GPT 5.5 (high) | OpenAI | 73.3% | $3.77 | 35 |
| 3 | Claude Opus 4.6 | Anthropic | 62.0% | $14.72 | 312 |
| 4 | Claude Sonnet 4.6 | Anthropic | 57.6% | $20.85 | 417 |
| 5 | GPT 5.5 | OpenAI | 48.8% | $1.46 | 17 |
| 6 | Claude Opus 4.7 (xhigh) | Anthropic | 33.5% | $7.67 | 169 |
| 7 | GPT 5.4 | OpenAI | 32.9% | $0.22 | 8 |
| 8 | Claude Opus 4.7 | Anthropic | 32.2% | $2.28 | 84 |
| 9 | Gemini 3.1 Pro | 27.1% | $3.32 | 161 | |
| 10 | Claude Haiku 4.5 | Anthropic | 19.3% | $1.04 | 133 |
| 11 | GPT 5.4 mini | OpenAI | 12.0% | $0.05 | 31 |
| 12 | GPT 5 mini | OpenAI | 10.1% | $0.03 | 19 |
| 13 | Gemini 3 Flash | 6.8% | $0.73 | 127 |