429generated behavioral tests
74.6%official best score
71.6%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) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 71.6% | ok | 307/429 | n/a | 82 | 0.25h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 69.9% | ok | 300/429 | $3.88 | 71 | 0.22h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 62.2% | ok | 267/429 | n/a | 44 | 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) · 20260519T215046Z
Public eval evidence has not been exported for this row yet.
GPT 5.5 (xhigh) · 20260517T094700Z
69.9% from 300/429 ProgramBench-scored tests. Raw public eval statuses: failure: 131, passed: 309.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
e592c32aec70eval.tests.test_basic.TestSeqCommand.test_seq_nonexistent_file (failure)e592c32aec70eval.tests.test_basic.TestSeqCommand.test_seq_strip_whitespace (failure)e592c32aec70eval.tests.test_basic.TestSeqCommand.test_seq_fake_quality (failure)e592c32aec70eval.tests.test_basic.TestSeqCommand.test_seq_x_lowercase_to_n (failure)e592c32aec70eval.tests.test_basic.TestSeqCommand.test_seq_reverse_both_strands (failure)e592c32aec70eval.tests.test_comp.TestCompCommand.test_comp_with_region_bed (failure)
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 | 74.6% | $8.12 | 77 |
| 2 | Claude Opus 4.7 (xhigh) | Anthropic | 73.4% | $9.95 | 147 |
| 3 | GPT 5.5 (high) | OpenAI | 68.5% | $3.38 | 46 |
| 4 | Claude Opus 4.6 | Anthropic | 67.4% | $14.11 | 306 |
| 5 | Claude Sonnet 4.6 | Anthropic | 64.8% | $19.13 | 482 |
| 6 | Claude Opus 4.7 | Anthropic | 63.9% | $2.48 | 68 |
| 7 | GPT 5.5 | OpenAI | 51.3% | $0.94 | 18 |
| 8 | GPT 5.4 | OpenAI | 40.3% | $0.26 | 12 |
| 9 | Gemini 3 Flash | 21.4% | $0.24 | 92 | |
| 10 | Gemini 3.1 Pro | 20.0% | $1.26 | 125 | |
| 11 | Claude Haiku 4.5 | Anthropic | 17.0% | $0.57 | 84 |
| 12 | GPT 5 mini | OpenAI | 8.9% | $0.01 | 16 |
| 13 | GPT 5.4 mini | OpenAI | 6.8% | $0.02 | 9 |