Human Frontier
87.7
Human-calibrated frontier signal, backed by Arena-style preference evidence and separate from raw AgMoBench benchmark composite scores.
Blended Price
Free/M
Input Price
Free/M
Output Price
Free/M
Speed
—
TTFT
—
Axes
Filters
Providers
vcontinuous-leaderboard · tasks-created:2026-02-01:2026-05-15 · gpt-5.4-2026-03-05-medium · harness swe-rebench-react@tools · effort medium · 5 attempts · pass@1 · swe-rebench:standardized:128k-or-model-limit · provider gpt-5.4-2026-03-05-medium__tools · record gpt-5.4-2026-03-05-medium__tools:1769904000000:1778803200000
{
"passAt5": 70.65868263473054,
"developer": "openai",
"instanceType": "model",
"contextPolicy": "128k unless the model supports less",
"attemptsPerTask": 5,
"totalTokenUsage": 814025.7628742515,
"modelReleaseDate": "2026-03-05",
"scoreStandardError": 0.8122550877994328,
"cachedTokenPercentage": 81.4876775099086,
"modelReleaseTimestamp": 1772668800000,
"standardContextTokens": 128000,
"averageInstanceCostUsd": 0.6071112652694611,
"potentialContamination": true,
"aggregateTaskRangeTimestamp": {
"to": 1778803200000,
"from": 1769904000000
},
"publishedTaskRangeTimestamp": {
"to": 1778803200000,
"from": 1769904000000
}
}Evidence observed 2026-05-15