Skip to content
aiqbench

Ant-Group · model passport

Ling Flash 2.0

ling-flash-2-0
  • Open weights
  • MIT licence
  • 102.9B params
  • Released 2025-09-17(HF repo created)
7/27axes covered
Overall · LMArena
1,343#210 of 396
Best percentile
62.2LiveCodeBench (Artificial Analysis)
Cohort standing
49.0median 52.6

What the data says

Benchmarked on 7 of 27 tracked axes. Every figure below is the model's own published result on that board.

Uneven: 35.3 percentile points between its best and worst axis — 62.2 pct on LiveCodeBench (Artificial Analysis) versus 26.9 pct on IFBench (Artificial Analysis).

  • LiveCodeBench (Artificial Analysis)62.2 pct58.9%Verify →
  • MMLU-Pro (Artificial Analysis)58.8 pct77.7%Verify →
  • Coding52.2 pct1,409Verify →
  • Hard prompts49.2 pct1,365Verify →
  • Overall47.2 pct1,343Verify →
  • Instruction following43.9 pct1,316Verify →
  • IFBench (Artificial Analysis)26.9 pct34.4%Verify →
  • 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
  • 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.

How to run it

Run it yourself

RepositoryinclusionAI/Ling-flash-2.0
Parameters102.9B
ArchitectureBailingMoeV2ForCausalLM
LicenceMIT

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "inclusionAI/Ling-flash-2.0"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Ollama

We don't sync Ollama tags, so we won't print a command we can't verify — a wrong one would pull the wrong model. Search the Ollama library →

Unsloth fine-tuning recipes are model-specific and we don't derive them here. Unsloth docs →

Identifiers, limits and prices are synced from the upstream source shown above. Any field we haven't synced is left out entirely rather than guessed.

  • The ling-flash-2.0 model from ant-group has a shrunk composite score of 49 (raw composite 48.6), with scores on 7 of 27 axes, and its coverage is not flagged as thin.

    standingVerify →

  • The model ling-flash-2.0 from ant-group has a standing of 49, compared with a cohort median standing of 52.6 across a cohort size of 147 over a cohort window of 180 days.

    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 →