Skip to content
aiqbench

DeepSeek · model passport

DeepSeek: R1 Distill Llama 70B

deepseek-r1-distill-llama-70b
  • Open weights
  • 70.6B params
  • 8,192 context
  • Released 2025-01-20(curated)
6/27axes covered
Best percentile
64.8MMLU-Pro (Artificial Analysis)
Cohort standing
43.3median 42.5

What the data says

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

Weakest axes

  • 13.3 pct — IFBench (Artificial Analysis) (27.6%)

Uneven: 51.5 percentile points between its best and worst axis — 64.8 pct on MMLU-Pro (Artificial Analysis) versus 13.3 pct on IFBench (Artificial Analysis).

  • MMLU-Pro (Artificial Analysis)64.8 pct79.5%Verify →
  • LiveBench52.6 pct54.7%Verify →
  • GPQA Diamond38.5 pct55.7%Verify →
  • AIME (math)38.5 pct51.4%Verify →
  • LiveCodeBench (Artificial Analysis)32.4 pct26.6%Verify →
  • IFBench (Artificial Analysis)13.3 pct27.6%Verify →
  • OverallNot benchmarked
  • WebDev ArenaNot benchmarked
  • Vision ArenaNot benchmarked
  • Factuality ArenaNot benchmarked
  • Search ArenaNot benchmarked
  • Document ArenaNot benchmarked
  • Agent ArenaNot 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
  • 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-R1-Distill-Llama-70B
Parameters70.6B
ArchitectureLlamaForCausalLM

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "deepseek-ai/DeepSeek-R1-Distill-Llama-70B"
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

API model IDdeepseek/deepseek-r1-distill-llama-70b
Context length8,192 tokens
Max output7,372 tokens
Input price$0.80 per 1M tokens
Output price$0.80 per 1M tokens

API parameters this endpoint supports

Frequency penaltyInclude reasoningMax tokensPresence penaltyReasoningRepetition penaltySeedStop sequencesTemperatureTop-kTop-p

Example request

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek/deepseek-r1-distill-llama-70b","messages":[{"role":"user","content":"Hello"}]}'

View on OpenRouter →

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-R1-Distill-Llama-70B stands at 43.3 (raw mean 40) — measured on 6 of 27 axes.

    standingtemplateVerify →

  • DeepSeek-R1-Distill-Llama-70B stands at 43.3 versus a cohort median of 42.5 across 141 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 →