2026-09-22T16:03:50.148Z
- performance:
{"outputTokensPerSecond":672.068,"timeToFirstAnswerTokenSeconds":5.56,"timeToFirstTokenSeconds":5.56}→ {"outputTokensPerSecond":746.487,"timeToFirstAnswerTokenSeconds":4.366,"timeToFirstTokenSeconds":4.366}
SOURCE-LINKED INTELLIGENCE
Mercury 2: terminalbench hard
Benchmark data by Artificial Analysis. Current measurement; upstream does not provide a measurement timestamp or benchmark revision here.
Read original source ↗ Open in workspace
- recordType
- observation
- region
- Global
- value
- 0.265151515151515
- unit
- score
- dimensions
- {"metric":"terminalbench_hard","model":"016d330a-2141-4afa-b2fc-62b314423dc1","parallel_queries":1,"prompt_length":1000}
Evidence & attribution
Artificial Analysis
First collected: 2026-09-19T21:50:47.018Z. This is not the publication date.
Observed changes
AIIC observation times, not verified publisher revision times. Up to eight recent revisions.
2026-09-20T21:52:07.471Z
- pricing:
Not provided→ {"input":0.25,"output":0.75} - performance:
Not provided→ {"outputTokensPerSecond":672.068,"timeToFirstAnswerTokenSeconds":5.56,"timeToFirstTokenSeconds":5.56}