1425generated behavioral tests
22.6%official best score
32.1%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 | 32.1% | ok | 458/1425 | $4.10 | 42 | 0.23h | manifest · eval json · eval summary · usage audit |
| 2 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 22.5% | ok | 321/1425 | n/a | 40 | 0.21h | not exported |
| 3 | GPT 5.5 (xhigh) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 13.5% | ok | 192/1425 | n/a | 45 | 0.21h | 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
32.1% from 458/1425 ProgramBench-scored tests. Raw public eval statuses: error: 2, failure: 1257, passed: 551, skipped: 9.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch.
60fc4d786b87tests.test_basic.TestBasicInvocation.test_unknown_command (failure)60fc4d786b87tests.test_basic.TestBasicInvocation.test_no_args_shows_help (failure)60fc4d786b87tests.test_coverage_gap.TestDepadCommand.test_depad_basic (failure)60fc4d786b87tests.test_coverage_gap.TestConsensusCommand.test_consensus_fastq_format (failure)60fc4d786b87tests.test_coverage_gap.TestConsensusCommand.test_consensus_all_bases (failure)60fc4d786b87tests.test_coverage_gap.TestStatsAdvanced.test_stats_coverage_distribution (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 | 22.6% | $11.21 | 70 |
| 2 | GPT 5.5 (high) | OpenAI | 21.5% | $5.82 | 35 |
| 3 | GPT 5.5 | OpenAI | 15.6% | $1.33 | 15 |
| 4 | Claude Opus 4.6 | Anthropic | 14.2% | $10.17 | 192 |
| 5 | GPT 5.4 | OpenAI | 7.1% | $0.33 | 10 |
| 6 | Gemini 3.1 Pro | 5.3% | $1.43 | 99 | |
| 7 | Gemini 3 Flash | 3.9% | $0.34 | 97 | |
| 8 | Claude Opus 4.7 (xhigh) | Anthropic | 3.9% | $1.81 | 56 |
| 9 | GPT 5.4 mini | OpenAI | 3.9% | $0.04 | 8 |
| 10 | Claude Opus 4.7 | Anthropic | 3.7% | $0.81 | 34 |
| 11 | Claude Haiku 4.5 | Anthropic | 3.0% | $0.24 | 67 |
| 12 | GPT 5 mini | OpenAI | 1.8% | $0.02 | 16 |
| 13 | Claude Sonnet 4.6 | Anthropic | 1.3% | $26.62 | 457 |