Human Frontier
85.3
Human-calibrated frontier signal, backed by Arena-style preference evidence and separate from raw AgMoBench benchmark composite scores.
Blended Price
$0.32/M
Input Price
$0.28/M
Output Price
$0.42/M
Speed
—
TTFT
—
Axes
Filters
Providers
vcontinuous-leaderboard · tasks-created:2025-11-01:2026-03-01 · DeepSeek-V3.2 · harness swe-rebench-react@tools · 5 attempts · pass@1 · swe-rebench:standardized:128k-or-model-limit · provider DeepSeek-V3.2__tools · record DeepSeek-V3.2__tools:1761955200000:1772323200000
{
"passAt5": 65,
"developer": "deepseek",
"instanceType": "model",
"contextPolicy": "128k unless the model supports less",
"attemptsPerTask": 5,
"totalTokenUsage": 2349749.462790721,
"modelReleaseDate": "2025-12-01",
"scoreStandardError": 0.763306626834098,
"cachedTokenPercentage": null,
"modelReleaseTimestamp": 1764547200000,
"standardContextTokens": 128000,
"averageInstanceCostUsd": 0.44236133086516966,
"potentialContamination": true,
"aggregateTaskRangeTimestamp": {
"to": 1772323200000,
"from": 1761955200000
},
"publishedTaskRangeTimestamp": {
"to": 1772323200000,
"from": 1761955200000
}
}Evidence observed 2026-03-01