1043generated behavioral tests
46.8%official best score
45.8%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 | 45.8% | ok | 478/1043 | n/a | 30 | 0.13h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 42.3% | ok | 441/1043 | $4.94 | 75 | 0.20h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 13.6% | ok | 142/1043 | n/a | 56 | 0.24h | 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
42.3% from 441/1043 ProgramBench-scored tests. Raw public eval statuses: failure: 626, passed: 458, skipped: 1.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
4ea5f4413e34eval.tests.test_7zip.test_add_with_timing_bt (failure)4ea5f4413e34eval.tests.test_7zip.test_add_split_volumes (failure)4ea5f4413e34eval.tests.test_7zip.test_add_with_log_level_bb1 (failure)4ea5f4413e34eval.tests.test_7zip.test_list_password_protected_archive (failure)4ea5f4413e34eval.tests.test_7zip.test_add_with_password (failure)4ea5f4413e34eval.tests.test_7zip.test_list_with_technical_info_slt (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 (high) | OpenAI | 46.8% | $4.19 | 46 |
| 2 | GPT 5.5 (xhigh) | OpenAI | 41.7% | $7.05 | 59 |
| 3 | Claude Sonnet 4.6 | Anthropic | 33.9% | $19.38 | 343 |
| 4 | GPT 5.5 | OpenAI | 28.8% | $1.01 | 13 |
| 5 | Gemini 3.1 Pro | 14.9% | $1.39 | 106 | |
| 6 | GPT 5.4 | OpenAI | 9.9% | $0.31 | 9 |
| 7 | Claude Opus 4.6 | Anthropic | 5.2% | $6.02 | 218 |
| 8 | Gemini 3 Flash | 4.8% | $0.19 | 65 | |
| 9 | Claude Opus 4.7 (xhigh) | Anthropic | 2.8% | $2.95 | 62 |
| 10 | GPT 5 mini | OpenAI | 2.4% | $0.02 | 6 |
| 11 | GPT 5.4 mini | OpenAI | 2.2% | $0.04 | 7 |
| 12 | Claude Opus 4.7 | Anthropic | 1.2% | $0.60 | 35 |
| 13 | Claude Haiku 4.5 | Anthropic | 0.0% | $1.10 | 157 |