Automated page. Written and translated automatically by AI and checked against the sources by software; no human editor. SI-IQ is an index built from public benchmark results, not an IQ test. How it works

How smart is DeepSeek-V4-Pro?

117 SI-IQ · 90% interval 113–120
Rank 46 of 80 ranked models · Plausible rank range 34–62

DeepSeek-V4-Pro is a model from DeepSeek, released on Apr 24, 2026. Its SI-IQ is 117 (90% interval 113–120), estimated from 9 benchmark results by 2 independent evaluators. Its strongest area is Knowledge and its weakest is Reasoning.

Facts

Developer: DeepSeek
Released: Apr 24, 2026
Access: Open weights (restricted licence)
Benchmarks used: 9 results from 2 evaluators
API price: $1.32 input / $3.96 output per 1M tokens (developer's price list, Oct 18, 2026)
Context window: 1,000,000 tokens

Ability by area (SI-IQ, 90% interval)

Every benchmark result we used

Each score is one value exactly as the source published it: for Epoch AI the best of its run settings, for OpenRouter the median hosting provider (the developer's own hosting when it ran the test). “Implied SI-IQ” is the ability that this single result points to; the overall SI-IQ is their precision-weighted average.

Benchmark	Area	Score	Implied SI-IQ	Evaluated by	As listed by the source	Retrieved	Source
Chess Puzzles	Reasoning	19.0%	111	Epoch AI	`deepseek-v4-pro_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
Mystery Game Puzzles	Reasoning	17.0%	109	Epoch AI	`deepseek-v4-pro_none`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
FrontierMath Tier 4	Math	2.4%	109	Epoch AI	`deepseek-v4-pro_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
FrontierMath Tiers 1–3	Math	45.3%	118	Epoch AI	`deepseek-v4-pro_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
OTIS Mock AIME 2024–2025	Math	96.7%	127	Epoch AI	`deepseek-v4-pro_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
SWE-bench Verified (Epoch AI run)	Coding	77.6%	126	Epoch AI	`deepseek-v4-pro_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
GPQA Diamond	Knowledge	90.9%	126	Epoch AI	`deepseek-v4-pro_high`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
GPQA Diamond (OpenRouter run)	Knowledge	85.4%	120	OpenRouter	`deepseek/deepseek-v4-pro-20260423` DeepSeek · first-party provider (1 of 21 rows)	Oct 18, 2026	OpenRouter benchmark datasets · data file
SimpleQA Verified	Knowledge	47.0%	126	Epoch AI	`deepseek-v4-pro_max`	Oct 18, 2026	Epoch AI Benchmarking Hub · data file
τ²-bench Airline (OpenRouter run)	Agents	78.9%	137	OpenRouter	`deepseek/deepseek-v4-pro-20260423` DeepSeek · first-party provider (1 of 20 rows)	Oct 18, 2026	OpenRouter benchmark datasets · data file

Human preference (Arena)

On the Arena leaderboard, where people vote between two anonymous answers, DeepSeek-V4-Pro has a rating of 1,458 (1,454–1,462) from 57,594 votes (published Oct 15, 2026). Human preference is a different measure from benchmark ability and is not part of SI-IQ. Arena (Hugging Face)

Models with similar SI-IQ

Model	SI-IQ
GPT-5.4 MiniOpenAI	117	DeepSeek-V4-Pro vs GPT-5.4 Mini
o3OpenAI	116	DeepSeek-V4-Pro vs o3
GLM-5.1Z.ai (Zhipu AI)	116	DeepSeek-V4-Pro vs GLM-5.1
GPT-5.4 NanoOpenAI	118	DeepSeek-V4-Pro vs GPT-5.4 Nano
Inkling-SmallThinking Machines	118	DeepSeek-V4-Pro vs Inkling-Small

Computed with SI-IQ method version 1.0.0. Read the full methodology

Data sources: Arena leaderboard dataset on Hugging Face (CC BY 4.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); API list prices via LiteLLM's price table (MIT); each price links to the developer's pricing page; Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter..