BenchLeader
Menu
LeaderboardsModelsProvidersBenchmarksCompareReleasesBlogMethodology
Hugging FaceAuto-detected

Zephyr 7B

Not yet ranked: too few independent results so far.

Blended price
–
Output speed
–
First answer
–
Context
–
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Reasoning24
  2. Coding25
  3. Human preference26

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 Prompts1116#328LMArena

Coding

BenchmarkdefaultSource
LMArena Coding1153#323LMArena

Human preference

BenchmarkdefaultSource
LMArena Text1130#324LMArena

See also

All Hugging Face models

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