---
title: "DeepSeek R1 671B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for DeepSeek R1."
source: "https://rankedagi.com/models/deepseek-r1"
---

# DeepSeek R1

| Field | Value |
| --- | --- |
| Organization | DeepSeek |
| License | Open Source |
| Size | 671B |
| Version | Latest |
| Released | 2025-01-20 |
| Context window | 128K |
| Knowledge cutoff | 2024-07-01 |
| Input cost per million tokens | $0.55 |
| Output cost per million tokens | $2.19 |
| Cached input cost per million tokens | $0.14 |
| Model ID | deepseek-reasoner |
| Last updated | 2026-05-04T01:36:17.118Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 47.2% | RankedAGI Coding Score |  |
| SWEBench Verified | coding | 49.2% | Agentic Coding | https://www.swebench.com/index.html |
| RankedAGI Agentic | Agents | 38.9% | RankedAGI Agentic Score |  |
| LiveCodeBench v5 | coding | 64.3% | LiveCodeBench - Evaluation of Large Language Models for Code |  |
| ChatArena (LMSYS) | coding | 1342 | ChatArena (LMSYS) Coding ELO Score with Style Control |  |
| Codeforces ELO Score | coding | 2029 | Codeforces ELO Score |  |
| Aider Polyglot | coding | 56.9% | Aider Polyglot Code Completion Benchmark |  |
| LiveBench Coding Score (old) | coding | 66.7% | LiveBench Coding Score old version |  |
| RankedAGI Reasoning | reasoning | 43.9% | RankedAGI Reasoning Score |  |
| Humanity's Last Exam | reasoning | 8.6% | Multidisciplinary Reasoning (no tools) |  |
| GPQA Diamond | reasoning | 71.5% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| Text Arena | general | 1398 | ChatArena (LMSYS) ELO Score | https://arena.ai/leaderboard/text |
| AIME 2025 | math | 70.0% | AIME 2025 Competition Math |  |
| NYT Connections Extended | reasoning | 38.6% | NYT Connections Extended Version | https://github.com/lechmazur/nyt-connections/ |
| MMLU Pro | knowledge | 84.0% | Graduate-level visual problem-solving |  |
| MMLU | knowledge | 90.8% | Image Reasoning |  |
| Hughes Hallucination Evaluation | safety | 14.3% | Computed using Vectara's Hallucination Evaluation Model, also known as HHEM. This evaluates how often an LLM introduces hallucinations when summarizing a document. | https://github.com/vectara/hallucination-leaderboard |
| LiveBench Average | general | 71.4% | LiveBench Average Score old version |  |
| Instruction Following Evaluation | instruction-following | 83.3% | Instruction Following Evaluation |  |
| LiveBench Coding 25.4 | coding | 48.5% | LiveBench Coding Score 2025-04 | https://livebench.ai/ |
| LiveBench Data Analysis | general | 69.8% | LiveBench Data Analysis Score | https://livebench.ai/ |
| LiveBench Language | general | 48.5% | LiveBench Language Score | https://livebench.ai/ |
| Artificial Analysis Quality Index | general | 89 | Artificial Analysis Quality Index (third-party composite, deprecated) |  |
| RankedAGI Math | math | 62.6% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 46.9% | Overall RankedAGI score |  |
| GDPval-AA Elo | Agents | 1253 | Office Tasks (Artificial Analysis) |  |