615generated behavioral tests
88.0%official best score
81.0%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 | 81.0% | ok | 498/615 | $7.03 | 84 | 0.31h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 35.8% | ok | 220/615 | n/a | 63 | 0.19h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 33.2% | ok | 204/615 | n/a | 48 | 0.16h | 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
81.0% from 498/615 ProgramBench-scored tests. Raw public eval statuses: failure: 123, passed: 662.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
86ccb6017898eval.tests.test_configuration.test_config_path_invalid_toml (failure)86ccb6017898eval.tests.test_configuration.test_config_path_valid (failure)86ccb6017898eval.tests.test_configuration.test_config_with_custom_styles (failure)8d05e17648bbeval.tests.test_config.test_config_with_multiple_keywords (failure)8d05e17648bbeval.tests.test_follow_mode.test_follow_flag_short (failure)8d05e17648bbeval.tests.test_follow_mode.test_follow_and_print_flags_together (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 | 88.0% | $15.26 | 97 |
| 2 | GPT 5.5 (high) | OpenAI | 77.9% | $3.64 | 33 |
| 3 | Claude Sonnet 4.6 | Anthropic | 75.8% | $58.91 | 840 |
| 4 | Claude Opus 4.7 | Anthropic | 75.6% | $2.18 | 65 |
| 5 | Claude Opus 4.7 (xhigh) | Anthropic | 74.0% | $26.95 | 265 |
| 6 | Claude Opus 4.6 | Anthropic | 72.0% | $20.30 | 324 |
| 7 | GPT 5.5 | OpenAI | 68.5% | $1.45 | 24 |
| 8 | GPT 5.4 mini | OpenAI | 62.9% | $0.10 | 10 |
| 9 | GPT 5.4 | OpenAI | 59.7% | $0.26 | 8 |
| 10 | Gemini 3 Flash | 56.6% | $0.67 | 130 | |
| 11 | GPT 5 mini | OpenAI | 26.2% | $0.02 | 13 |
| 12 | Gemini 3.1 Pro | 2.1% | $1.56 | 51 | |
| 13 | Claude Haiku 4.5 | Anthropic | 0.0% | $0.87 | 101 |