DeepSeek · mit

DeepSeek-V4-Pro-Max

DeepSeek-V4-Pro-Max by DeepSeek appears in 1 source with Reasoning at 56.96. Best read for Code quality, Long context, Low cost.

CreatorDeepSeek

Release date2026-04-23

Knowledge cutoffNot published

Context1M tokens

Input price$1.74/M tokens

Output price$3.48/M tokens

Modalitytext

CountryCN

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank20ranking · source_rank

Reasoning56.96reasoning · index_reasoning

Math52.92math · index_math

Coding43.53coding · index_code

Research32.92research · index_search

Vision33.51multimodal · index_vision

Tool calling33.15tool_calling · index_tool_calling

Long context19.81long_context · index_long_context

Finance42.02domain · index_finance

Legal42.52domain · index_legal

Healthcare47.79domain · index_healthcare

GPQA90.1 %reasoning · gpqa_score

SWE-bench Verified80.6 %coding · swe_bench_verified_score

Code Arena1,317.2 %coding · coding_arena_score

Humanity Last Exam48.2 %reasoning · hle_score

SimpleQA57.9 %knowledge · simpleqa_score

BrowseComp83.4 %research · browsecomp_score

Toolathlon51.8 %tool_calling · toolathlon_score

MCP Atlas73.6 %tool_calling · mcp_atlas_score

SWE-bench Pro55.4 %coding · swe_bench_pro_score

Context1,048,576 tokenscontext · context

Speed43.9 c/sperformance · throughput

Latency126,951.07 msperformance · latency

Input price1.74 $/Mpricing · input_price

Output price3.48 $/Mpricing · output_price

Parameters1,600,000,000,000 paramsmodel · params

Evidence

Citations and source overlap

LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is DeepSeek-V4-Pro-Max verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.