Skip to content
aiqbench

DeepSeek · model passport

DeepSeek V4.1 Flash 1017

deepseek-v4-1-flash-1017
  • Open weights
  • 1,048,576 context
  • Released 2026-10-17(OpenRouter model id)
4/27axes covered
Best percentile
93.1AIME (math)
Cohort standing
66.3median 57.6

What the data says

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

Strongest axes

  • 93.1 pct — AIME (math) (99.4%)
  • 91.5 pct — GPQA Diamond (93.2%)

Uneven: 45.5 percentile points between its best and worst axis — 93.1 pct on AIME (math) versus 47.6 pct on SimpleQA Verified.

  • AIME (math)93.1 pct99.4%Verify →
  • GPQA Diamond91.5 pct93.2%Verify →
  • ARC-AGI-282.1 pct78.1%Verify →
  • SimpleQA Verified47.6 pct42.5%Verify →
  • OverallNot benchmarked
  • WebDev ArenaNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot benchmarked
  • FrontierMathNot 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
  • CodingNot benchmarked
  • Instruction followingNot benchmarked
  • Hard promptsNot 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-V4.1-Flash-1017

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "deepseek-ai/DeepSeek-V4.1-Flash-1017"
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 →

Use it hosted

ProviderDeepSeek
Context length1,048,576 tokens
Max output393,216 tokens
Acceptstext, image
Input price$0.15 per 1M tokens
Cached input price$0.00 per 1M tokens
Output price$0.60 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningLogit biasLogprobsMax completion tokensMax tokensMin-pParallel Tool CallsPresence penaltyReasoningReasoning effortRepetition penaltyResponse formatSeedStop sequencesStructured outputsTemperatureTool choiceTool callingTop-kTop logprobsTop-p

Official model notes →

Context, limits and prices come from the last catalogue sync; the API id, the request and the link come from the provider's own documentation. Any field we haven't synced is left out entirely rather than guessed.

  • The DeepSeek deepseek-v4.1-flash-1017 model covers 4 of 27 axes with a composite score of 78.6, resulting in a shrunk score of 66.3 and is not thin coverage.

    standingVerify →

  • The DeepSeek deepseek-v4.1-flash-1017 model achieved a standing of 66.3, which is above the cohort median standing of 57.6 across 224 models within an 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 →