1232generated behavioral tests
46.1%official best score
41.9%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 | 41.9% | ok | 516/1232 | $5.51 | 106 | 0.29h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 34.9% | ok | 430/1232 | n/a | 84 | 0.25h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 32.7% | ok | 403/1232 | n/a | 45 | 0.18h | 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
41.9% from 516/1232 ProgramBench-scored tests. Raw public eval statuses: failure: 767, passed: 879, skipped: 44.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, exact version/output mismatch.
b26d2e905b78tests.test_advanced_features.test_generate_with_complex_grammar (failure)b26d2e905b78tests.test_build.test_build_output_flag (failure)b26d2e905b78tests.test_build.test_build_verbose_flag (failure)b26d2e905b78tests.test_build.test_build_with_path_arg (failure)b26d2e905b78tests.test_advanced_features.test_build_with_different_flags (failure)b26d2e905b78tests.test_build.test_build_with_generated_parser (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 | 46.1% | $10.62 | 88 |
| 2 | GPT 5.5 (high) | OpenAI | 38.4% | $4.10 | 39 |
| 3 | Claude Opus 4.7 (xhigh) | Anthropic | 37.3% | $10.81 | 197 |
| 4 | Claude Sonnet 4.6 | Anthropic | 37.2% | $15.00 | 402 |
| 5 | Claude Opus 4.6 | Anthropic | 36.1% | $11.07 | 264 |
| 6 | Claude Opus 4.7 | Anthropic | 35.1% | $8.38 | 139 |
| 7 | GPT 5.5 | OpenAI | 34.2% | $1.04 | 19 |
| 8 | GPT 5.4 | OpenAI | 29.1% | $0.28 | 9 |
| 9 | Gemini 3.1 Pro | 23.0% | $0.89 | 73 | |
| 10 | Claude Haiku 4.5 | Anthropic | 14.0% | $0.53 | 121 |
| 11 | GPT 5 mini | OpenAI | 13.4% | $0.03 | 19 |
| 12 | Gemini 3 Flash | 6.9% | $0.30 | 90 | |
| 13 | GPT 5.4 mini | OpenAI | 6.2% | $0.05 | 74 |