---
title: "DeepSeek V4 Pro 1.6T Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for DeepSeek V4 Pro."
source: "https://rankedagi.com/models/deepseek-v4-pro"
---

# DeepSeek V4 Pro

| Field | Value |
| --- | --- |
| Organization | DeepSeek |
| License | Open Source |
| Type | Thinking |
| Size | 1.6T |
| Version | Latest |
| Released | 2026-04-24 |
| Context window | 1M |
| Input cost per million tokens | $0.44 |
| Output cost per million tokens | $0.87 |
| Cached input cost per million tokens | $0 |
| Last updated | 2026-05-28T08:57:27.471Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 77.5% | RankedAGI Coding Score |  |
| SWEBench Pro | coding | 55.4% | Diverse Agentic Coding |  |
| SWEBench Multilingual | coding | 76.2% |  |  |
| SWEBench Verified | coding | 80.6% | Agentic Coding | https://www.swebench.com/index.html |
| Terminal-Bench 2.0 | coding | 67.9% | Agentic Terminal Coding | https://www.tbench.ai/leaderboard/terminal-bench/2.0 |
| RankedAGI Agentic | Agents | 68.1% | RankedAGI Agentic Score |  |
| BrowseComp | Agents | 83.4% | A benchmark for browsing agents | https://openai.com/index/browsecomp/ |
| Code Arena | coding | 1464 | Code Arena Elo Score | https://arena.ai/leaderboard/code |
| Livebench Coding 26.1 | coding | 70.0% | Livebench Coding 26.01.08 | https://livebench.ai/#/?Coding=a&sort=Coding+Average&variants=true&highunseenbias=true |
| DeepSWE | coding | 8.0% | Measuring frontier coding agents on original, long-horizon engineering tasks | https://deepswe.datacurve.ai/ |
| LiveBench Agentic Coding 26.1 | coding | 56.7% |  | https://livebench.ai/#/?Agentic+Coding=a&sort=Agentic+Coding+Average&variants=true&highunseenbias=true |
| SvelteBench | coding | 91.1% | SvelteBench - Benchmark for Svelte | https://khromov.github.io/svelte-bench/benchmark-results-merged.html |
| LiveCodeBench v6 | coding | 93.5% | LiveCodeBench v6 - Evaluation of Large Language Models for Code |  |
| Codeforces ELO Score | coding | 3206 | Codeforces ELO Score |  |
| RankedAGI Reasoning | reasoning | 75.4% | RankedAGI Reasoning Score |  |
| Humanity's Last Exam | reasoning | 37.7% | Multidisciplinary Reasoning (no tools) |  |
| Humanity's Last Exam w/ Tools | reasoning | 48.2% | Multidisciplinary Reasoning (with tools) |  |
| GPQA Diamond | reasoning | 90.1% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| Text Arena | general | 1463 | ChatArena (LMSYS) ELO Score | https://arena.ai/leaderboard/text |
| Vending Bench 2 | Agents | $3,284.52 | Benchmark for measuring AI model performance on running a business over long time horizons. Models are tasked with running a simulated vending machine business over a year and scored on their bank account balance at the end. | https://andonlabs.com/evals/vending-bench-2 |
| MMLU Pro | knowledge | 87.5% | Graduate-level visual problem-solving |  |
| RankedAGI Math | math | 72.7% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 66.8% | Overall RankedAGI score |  |
| GDPval-AA Elo | Agents | 1554 | Office Tasks (Artificial Analysis) |  |
| MCP Atlas | Agents | 73.6% | Scaled Tool Use |  |
| Toolathlon | Agents | 51.8% | Benchmark to assess language agents' general tool use in realistic environments. | https://toolathlon.xyz/introduction |

## Model Links

| Title | URL |
| --- | --- |
| Release | https://api-docs.deepseek.com/news/news260424 |
| X | https://x.com/deepseek_ai/status/2047516922263285776 |