phi-4 vs Mistral Large

API pricing, benchmark indices and specs — updated 19 October 2026.

microsoft · A · Open weights
Params
14.7 B
Context
16,384
30d
356 K
mit · commercial OK
mistralai · B · API-only
Input / 1M
$2.00
Output / 1M
$6.00
AA index
5.8
proprietary · 128K context

Specification comparison

Differences are highlighted; identical values are muted.

Specification phi-4 Mistral Large
Deployment Open weights · also via API API-only (closed weights)
Parameters 14.7 B n/a — API-only
Architecture Phi3ForCausalLM n/a — API-only
Context length 16,384 128,000
Licence mit Proprietary
Commercial use Allowed Per API terms
Languages 1 n/a — API-only
Downloads 30d 356,282 n/a — API-only
Downloads all time 14.8 M n/a — API-only
Quantizations on the Hub SAFETENSORS n/a — API-only
Gated No n/a — API-only
First seen on the Hub 2024-12-11 n/a — API-only
API input price / 1M tokens $0.07 $2.00
API output price / 1M tokens $0.14 $6.00
Est. cost per 1M input + 1M output tokens $0.21 $8.00
Cached input / 1M tokens — $0.20
AA Intelligence Index 5.9 5.8
AA Coding Index — —
AA Agentic Index — —
Max output tokens 14,745 102,400
API providers 1 1
Tool calling No Yes
Reasoning mode No No

API cost

Estimated cost of 1M input + 1M output tokens via OpenRouter

phi-4 $0.21
Mistral Large $8.00

phi-4 is 38.1× cheaper at list price.

VRAM side by side

On RTX 4090 · 24 GB · 8K context unless noted

phi-4 · fp16 34.9 GB / 24 GB
❌ Won’t fit

Adjacent comparisons