1586generated behavioral tests
85.2%official best score
85.1%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) | 20260518T232006Z |
Codex /goal | Paper prompt + /goal no internet | 85.1% | ok | 1350/1586 | n/a | 60 | 0.16h | not exported |
| 2 | GPT 5.5 (xhigh) | 20260517T094700Z |
Codex /goal | Mini-SWE-compatible no internet | 84.1% | ok | 1334/1586 | $3.08 | 62 | 0.17h | manifest · eval json · eval summary · usage audit |
| 3 | GPT 5.5 (xhigh) | 20260519T215046Z |
Codex /goal | Paper prompt + Goal contract no internet | 83.9% | ok | 1330/1586 | n/a | 45 | 0.12h | 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
84.1% from 1334/1586 ProgramBench-scored tests. Raw public eval statuses: error: 1, failure: 304, passed: 1924, skipped: 8.
Agent trace summary unavailable.
Likely miss class: behavioral mismatch, interactive terminal behavior mismatch, terminal rendering mismatch.
8ec0cc6e2379tests.test_tmux_fixed.test_tmux_log_view (failure)8ec0cc6e2379tests.test_tmux_fixed.test_tmux_show_view (failure)f276bd1d6c4eeval.tests.test_env_and_config.test_tigrc_user_custom_location (failure)f276bd1d6c4eeval.tests.test_env_and_config.test_tigrc_user_nonexistent (failure)f276bd1d6c4eeval.tests.test_env_and_config.test_tigrc_system_custom_location (failure)f276bd1d6c4eeval.tests.test_env_and_config.test_tig_diff_opts (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 (xhigh) | OpenAI | 85.2% | $4.82 | 39 |
| 2 | GPT 5.5 (high) | OpenAI | 85.0% | $2.06 | 32 |
| 3 | GPT 5.5 | OpenAI | 84.2% | $0.82 | 16 |
| 4 | Gemini 3.1 Pro | 83.9% | $1.02 | 72 | |
| 5 | Claude Opus 4.7 (xhigh) | Anthropic | 83.8% | $1.57 | 65 |
| 6 | Claude Opus 4.7 | Anthropic | 83.6% | $0.84 | 44 |
| 7 | Claude Haiku 4.5 | Anthropic | 82.9% | $0.22 | 80 |
| 8 | Claude Opus 4.6 | Anthropic | 79.6% | $10.78 | 281 |
| 9 | GPT 5 mini | OpenAI | 76.9% | $0.01 | 12 |
| 10 | Gemini 3 Flash | 71.2% | $0.19 | 54 | |
| 11 | GPT 5.4 mini | OpenAI | 39.0% | $0.03 | 8 |
| 12 | GPT 5.4 | OpenAI | 38.1% | $0.12 | 7 |
| 13 | Claude Sonnet 4.6 | Anthropic | 0.0% | $14.93 | 382 |