gpt-5.4-mini-2026-03-17 vs mistral-small-2603

Pricing, Performance & Features Comparison

gpt-5.4-mini-2026-03-17

Authoropenai

Context Length400K

Reasoning

Providers1

ReleasedMar 2026

Knowledge CutoffAug 2025

License–

Input$0.75

Output$4.5

Latency (p50)4.9s

Output Limit128K

Function Calling

JSON Mode

InputText, Image

OutputText

openai

in$0.75out$4.5cache$0.075–

Latency (24h)

Success Rate (24h)

mistral-small-2603

Authormistral

Context Length256K

Reasoning

Providers1

ReleasedMar 2026

Knowledge Cutoff–

License–

Mistral Small 4: a hybrid Apache 2.0 model consolidating Magistral reasoning, Devstral coding agents, and Mistral Small into one model with toggleable reasoning effort.

Input$0.15

Output$0.6

Latency (p50)3.8s

Output Limit–

Function Calling

JSON Mode

InputText, Image

OutputText

mistral

in$0.15out$0.6––