BenchLeader
Prime IntellectAuto-detected

Intellect 3

Best configuration ranks #360 of 610 on the BenchLeader Index at 47.0 ±4.4. Last measured 2 Sept 2026.

Blended price
Output speed
First answer
Context
131k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index47
  2. Reasoning53
  3. Coding54
  4. Agents & tools42
  5. Knowledge40
  6. Instruction following37
  7. Human preference53
  8. Long context45
  9. Composite42

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

BenchmarkdefaultSource
LMArena Hard Prompts1374#173LMArena
GPQA Diamond (AA)not in index76.1%#232Artificial Analysis
Humanity's Last Exam (AA)not in index13.1%#212Artificial Analysis

Coding

BenchmarkdefaultSource
LMArena Coding1412#164LMArena

Agents & tools

Knowledge

BenchmarkdefaultSource
AA-Omniscience-50.9#324Artificial Analysis

Instruction following

BenchmarkdefaultSource
IFBench34.0%#308Artificial Analysis

Human preference

BenchmarkdefaultSource
LMArena Text1356#173LMArena

Long context

BenchmarkdefaultSource
AA-LCR39.7%#289Artificial Analysis

Composite

BenchmarkdefaultSource
AA Intelligence Index10.6#303Artificial Analysis

See also

Data as of 9 Sept 2026. Compare with another model.