LLM ReferenceLLM Reference
All comparisons

GPT-5.4 Mini vs DeepSeek V3

Side-by-side comparison of specifications, capabilities, and pricing.

Released2026-03-052024-12-26
Context window400K64k
Parameters671B
Architecturedecoder onlymixture of experts
LicenseProprietaryOpen Source
Knowledge cutoff2025-082024-04

Capabilities

Vision
Multimodal
Reasoning
Function calling
Tool use
Structured Outputs
Code execution

Availability

Providers

Benchmarks

HellaSwag—95.7HumanEval—85.5Massive Multitask Language Understanding—88.5LiveCodeBench—49.6Aider Polyglot—48.4BigCodeBench—50.0Chatbot Arena—1302.0MMLU PRO—75.9