Skip to content
aiqbench

Model passport

Koala 13b

koala-13b
  • Open weights
  • Non-commercial licence
  • Released 2023-04-03(curated)
4/27axes covered
Overall · LMArena
1,071#384 of 396
Best percentile
3.3Overall
Cohort standing
23.1median 25.6

What the data says

Benchmarked on 4 of 27 tracked axes. Every figure below is the model's own published result on that board.

Weakest axes

  • 2.8 pct — Hard prompts (1,035)
  • 2.8 pct — Instruction following (1,027)
  • 2.8 pct — Coding (1,066)

Uneven: 0.5 percentile points between its best and worst axis — 3.3 pct on Overall versus 2.8 pct on Hard prompts.

  • Overall3.3 pct1,071Verify →
  • Coding2.8 pct1,066Verify →
  • Instruction following2.8 pct1,027Verify →
  • Hard prompts2.8 pct1,035Verify →
  • WebDev ArenaNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • GPQA DiamondNot benchmarked
  • AIME (math)Not benchmarked
  • FrontierMathNot benchmarked
  • SimpleQA VerifiedNot benchmarked
  • ARC-AGI-2Not benchmarked
  • Humanity's Last ExamNot benchmarked
  • SimpleBenchNot benchmarked
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not benchmarked
  • MMLU-Pro (Artificial Analysis)Not benchmarked
  • LiveCodeBench (Artificial Analysis)Not benchmarked
  • IFBench (Artificial Analysis)Not benchmarked
  • HHEM Hallucination (Vectara)Not benchmarked
  • SWE-bench VerifiedNot benchmarked
  • Terminal-BenchNot benchmarked

Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.

  • The koala-13b model has a shrunk composite score of 23.1 and a raw composite score of 2.9, with coverage on 4 of 27 axes.

    standingVerify →

  • koala-13b has a standing of 23.1, compared with a cohort median standing of 25.6 across a cohort of 29 models over a 180-day window.

    cohortVerify →

Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →