1161generated behavioral tests
78.9%official best score
72.9%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) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 72.9% | ok | 846/1161 | $4.50 | 83 | 0.25h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 72.5% | ok | 842/1161 | n/a | 57 | 0.23h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 69.1% | ok | 802/1161 | n/a | 54 | 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) · 20260517T094700Z
72.9% from 846/1161 ProgramBench-scored tests. Raw public eval statuses: error: 1, failure: 323, passed: 1203, skipped: 26.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
9a09132872e3eval.tests.test_additional_coverage.test_log_file (failure)9a09132872e3eval.tests.test_config_and_sorting.test_log_file_gets_written (failure)9a09132872e3eval.tests.test_config_comprehensive.test_config_archive_browsing (failure)9a09132872e3eval.tests.test_coverage_boost.test_log_file_with_content (failure)9a09132872e3eval.tests.test_coverage_improvement.test_single_file_scan (failure)9a09132872e3eval.tests.test_additional_coverage.test_database_with_filtering (failure)
GPT 5.5 (xhigh) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260518T232006Z
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 | 78.9% | $7.44 | 85 |
| 2 | GPT 5.5 (high) | OpenAI | 75.8% | $3.88 | 60 |
| 3 | Claude Sonnet 4.6 | Anthropic | 70.1% | $21.13 | 532 |
| 4 | Claude Opus 4.6 | Anthropic | 66.5% | $10.24 | 245 |
| 5 | GPT 5.5 | OpenAI | 64.2% | $1.02 | 18 |
| 6 | Claude Opus 4.7 (xhigh) | Anthropic | 63.2% | $5.45 | 96 |
| 7 | Claude Opus 4.7 | Anthropic | 59.6% | $2.35 | 69 |
| 8 | Claude Haiku 4.5 | Anthropic | 42.1% | $0.72 | 125 |
| 9 | Gemini 3.1 Pro | 36.3% | $1.17 | 105 | |
| 10 | GPT 5.4 mini | OpenAI | 32.4% | $0.03 | 7 |
| 11 | GPT 5.4 | OpenAI | 27.9% | $0.31 | 8 |
| 12 | Gemini 3 Flash | 9.5% | $0.28 | 73 | |
| 13 | GPT 5 mini | OpenAI | 4.5% | $0.01 | 6 |