Smaug
Smaug Agentic
newRank 5 of 45 · release 2026-06-25
Global average
79.5
first appearance
Signal breakdown
Category scores
Reasoning
90.3
Coding
82.5
Agentic Coding
64.7
Mathematics
83.9
Data Analysis
79.9
Language
84.4
Instruction Following
71.0
Trend
Score history
Analyst annotations
Commentary
Sign in to publish an annotation under your analyst identity.
No annotations yet
Be the first analyst on record for Smaug Agentic.
Detail
Per-task scores
| Task | Category | Score |
|---|---|---|
| spatial | 100.0 | |
| connections | 100.0 | |
| zebra_puzzle | 98.3 | |
| tablereformat | 98.0 | |
| AMPS_Hard | 98.0 | |
| math_comp | 94.1 | |
| olympiad | 91.6 | |
| consecutive_events | 90.4 | |
| code_generation | 84.5 | |
| logic_with_navigation | 84.0 | |
| typos | 82.0 | |
| code_completion | 80.4 | |
| theory_of_mind | 78.8 | |
| javascript | 77.3 | |
| story_generation | 76.7 | |
| summarize | 72.0 | |
| plot_unscrambling | 71.1 | |
| paraphrase | 68.6 | |
| simplify | 66.6 | |
| python | 60.0 | |
| typescript | 56.7 | |
| integrals_with_game | 52.0 | |
| tablejoin | 51.2 |