539generated behavioral tests
67.0%official best score
61.8%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) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 61.8% | ok | 333/539 | $4.80 | 52 | 0.17h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 60.9% | ok | 328/539 | n/a | 42 | 0.15h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 39.1% | ok | 211/539 | n/a | 26 | 0.14h | 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
61.8% from 333/539 ProgramBench-scored tests. Raw public eval statuses: error: 2, failure: 271, passed: 577.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, terminal rendering mismatch.
35a40756148ceval.tests.test_tui_advanced (error)35a40756148ceval.tests.test_tui_pexpect (error)92f620e1a32atests.test_config.test_write_default_conf_basic (failure)92f620e1a32atests.test_config_extended.test_write_default_conf_with_symlink_destination (failure)92f620e1a32atests.test_config.test_conf_flag_with_nonexistent_file (failure)92f620e1a32atests.test_config.test_conf_flag_with_malformed_file (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 | Claude Opus 4.6 | Anthropic | 67.0% | $14.42 | 241 |
| 2 | GPT 5.5 (xhigh) | OpenAI | 66.4% | $11.09 | 83 |
| 3 | GPT 5.5 | OpenAI | 62.3% | $1.89 | 24 |
| 4 | Claude Opus 4.7 | Anthropic | 60.9% | $1.57 | 58 |
| 5 | Claude Opus 4.7 (xhigh) | Anthropic | 59.0% | $6.43 | 140 |
| 6 | Claude Sonnet 4.6 | Anthropic | 56.4% | $23.30 | 406 |
| 7 | Gemini 3.1 Pro | 52.5% | $2.80 | 149 | |
| 8 | GPT 5.5 (high) | OpenAI | 51.9% | $4.81 | 46 |
| 9 | Gemini 3 Flash | 37.3% | $0.23 | 42 | |
| 10 | Claude Haiku 4.5 | Anthropic | 33.6% | $0.43 | 99 |
| 11 | GPT 5 mini | OpenAI | 31.4% | $0.01 | 9 |
| 12 | GPT 5.4 | OpenAI | 18.2% | $0.42 | 7 |
| 13 | GPT 5.4 mini | OpenAI | 7.2% | $0.08 | 28 |