Skip to content
aiqbench

DeepSeek · model passport

DeepSeek V2 5

deepseek-v2-5
  • Open weights
  • DeepSeek licence
  • 235.7B params
  • Released 2024-09-05(curated)
6/27axes covered
Overall · LMArena
1,307#257 of 396
Best percentile
41.2Coding
Cohort standing
41.3median 37.9

What the data says

Benchmarked on 6 of 27 tracked axes. Every figure below is the model's own published result on that board.

Uneven: 8.2 percentile points between its best and worst axis — 41.2 pct on Coding versus 33 pct on MMLU-Pro (Artificial Analysis).

  • Coding41.2 pct1,368Verify →
  • LiveBench38.8 pct48.1%Verify →
  • Hard prompts37.9 pct1,321Verify →
  • Instruction following35.9 pct1,291Verify →
  • Overall35.4 pct1,307Verify →
  • MMLU-Pro (Artificial Analysis)33 pct66.6%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
  • BFCL v4 (function calling)Not benchmarked
  • SWE-bench Verified (live board)Not benchmarked
  • Terminal-Bench 4.0Not 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.

How to run it

Run it yourself

Repositorydeepseek-ai/DeepSeek-V2.5
Parameters235.7B
ArchitectureDeepseekV2ForCausalLM
LicenceDeepSeek

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "deepseek-ai/DeepSeek-V2.5"
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-v2.5 stands at 41.3 (raw mean 37) — measured on 6 of 27 axes.

    standingtemplateVerify →

  • deepseek-v2.5 stands at 41.3 versus a cohort median of 37.9 across 130 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 →