mistral-large-2512 vs gpt-5.2-2025-12-11
Pricing, Performance & Features Comparison
Mistral Large 3: a 675B-parameter granular-MoE multimodal model released under Apache 2.0 as part of the Mistral 3 family.
Input$0.5
Output$1.5
Latency (p50)4.7s
Output Limit–
Function Calling
JSON Mode
-
InputText, Image
OutputText
in$0.5out$1.5––
Latency (24h)
Success Rate (24h)
OpenAI’s December 2025 frontier model, still generally available as the value option of the GPT-5.x line.
Input$1.8
Output$14
Latency (p50)4.4s
Output Limit128K
Function Calling
JSON Mode
-
InputText, Image
OutputText
in$1.8out$14cache$0.18–