Skip to content
aiqbench

DeepSeek · model passport

DeepSeek V3 1

deepseek-v3-1
  • Open weights
  • MIT licence
  • 684.5B params
  • Released 2025-08-21(curated)
9/27axes covered
Overall · LMArena
1,417#123 of 396
Best percentile
92.1MMLU-Pro (Artificial Analysis)
Cohort standing
63.1median 51.9

What the data says

Benchmarked on 9 of 27 tracked axes. Every figure below is the model's own published result on that board.

Strongest axes

  • 92.1 pct — MMLU-Pro (Artificial Analysis) (85.1%)
  • 88.5 pct — LiveCodeBench (Artificial Analysis) (78.4%)

Uneven: 59.9 percentile points between its best and worst axis — 92.1 pct on MMLU-Pro (Artificial Analysis) versus 32.2 pct on SimpleBench.

  • MMLU-Pro (Artificial Analysis)92.1 pct85.1%Verify →
  • LiveCodeBench (Artificial Analysis)88.5 pct78.4%Verify →
  • HHEM Hallucination (Vectara)83.8 pct94.5%Verify →
  • Overall69.2 pct1,417Verify →
  • Instruction following67.4 pct1,402Verify →
  • Hard prompts66.4 pct1,432Verify →
  • Coding63.2 pct1,447Verify →
  • IFBench (Artificial Analysis)44.4 pct41.5%Verify →
  • SimpleBench32.2 pct40.0%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
  • LiveBenchNot benchmarked
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not 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

Repositorydeepseek-ai/DeepSeek-V3.1
Parameters684.5B
ArchitectureDeepseekV3ForCausalLM
LicenceMIT

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "deepseek-ai/DeepSeek-V3.1"
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.

  • deepseek-v3.1 stands at 63.1 (raw mean 67.5) — measured on 9 of 27 axes.

    standingtemplateVerify →

  • deepseek-v3.1 stands at 63.1 versus a cohort median of 51.9 across 148 models released within 180 days — above the peer median.

    cohorttemplateVerify →

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 →