1235generated behavioral tests
82.8%official best score
81.6%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 | 81.6% | ok | 1008/1235 | $8.74 | 157 | 0.41h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 79.1% | ok | 977/1235 | n/a | 84 | 0.26h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 59.4% | ok | 734/1235 | n/a | 44 | 0.22h | 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
81.6% from 1008/1235 ProgramBench-scored tests. Raw public eval statuses: failure: 277, passed: 1118, skipped: 10.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
035fc64252a5eval.tests.test_completions.test_gen_completions_bash (failure)035fc64252a5eval.tests.test_completions.test_gen_completions_zsh (failure)035fc64252a5eval.tests.test_completions.test_gen_completions_fish (failure)035fc64252a5eval.tests.test_completions.test_gen_completions_powershell (failure)7af3fcb896a2tests.test_coverage_improvements.test_no_global_ignore_file (failure)7af3fcb896a2tests.test_coverage_improvements.test_max_buffer_time (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 | 82.8% | $8.84 | 94 |
| 2 | GPT 5.5 (high) | OpenAI | 80.2% | $3.69 | 55 |
| 3 | Claude Opus 4.7 | Anthropic | 78.1% | $3.02 | 80 |
| 4 | Claude Opus 4.7 (xhigh) | Anthropic | 76.4% | $8.46 | 143 |
| 5 | Claude Sonnet 4.6 | Anthropic | 74.8% | $17.57 | 384 |
| 6 | GPT 5.5 | OpenAI | 72.6% | $1.04 | 15 |
| 7 | Gemini 3.1 Pro | 54.5% | $1.69 | 111 | |
| 8 | Gemini 3 Flash | 48.9% | $0.26 | 88 | |
| 9 | GPT 5.4 | OpenAI | 39.8% | $0.28 | 12 |
| 10 | Claude Haiku 4.5 | Anthropic | 39.7% | $0.61 | 117 |
| 11 | GPT 5.4 mini | OpenAI | 3.6% | $0.03 | 8 |
| 12 | GPT 5 mini | OpenAI | 3.1% | $0.05 | 16 |
| 13 | Claude Opus 4.6 | Anthropic | 0.0% | $1.07 | 183 |