Who has the fastest Mistral Large 3 inference?

Groq serves Mistral Large 3 at 450 tokens/sec. Speed optimized. For latency-sensitive workloads this is usually the right pick even if not the cheapest.

Is the output quality identical across all providers hosting Mistral Large 3?

The weights are identical · Mistral Research License. Differences come from quantization (some providers use int8 or fp8 for speed), context window caps, and provider-added safety filters.

What are cheaper alternative models to Mistral Large 3?

See our substitute finder for models within 10% performance at lower price.

Arbitrage · Mistral Large 34 providers · $1.80 → $2.50 · 28% spread

Cheapest Provider for Mistral Large 3

Mistral's flagship · strong in code + multilingual · EU-hosted primary.

123BMistral Research LicenseModel detail page →

Cheapest input

Groq

$1.80/M

Speed optimized

Fastest

Groq

450 tok/s

Speed optimized

Savings calculator

Save 28%

vs Together AI at $2.50/M input. For 100M tokens/mo, that is $70/mo saved by routing to Groq.

Sorted by input price

All 4 providers

Provider	In $/M	Out $/M	Context	Speed	Free	Region
GroqWinner	$1.80	$5.40	128K	450 t/s		US
Mistral (direct)	$2.00	$6.00	128K	65 t/s		EU (Paris)
AA Azure AI	$2.00	$6.00	128K	70 t/s	—	Multi-region
Together AI	$2.50	$7.50	128K	85 t/s		US

Notes: Groq · Speed optimized ; Mistral (direct) · La Plateforme · EU hosted · GDPR ; Azure AI · Enterprise · SOC 2 · HIPAA ; Together AI · US hosted · faster

Frequently Asked Questions

Groq at $1.80/M input and $5.40/M output. That is 28% cheaper than Together AI. Speed optimized.

Cheapest Provider for Mistral Large 3

All 4 providers

Models cheaper than Mistral Large 3

Frequently Asked Questions

This model

Arbitrage

Explore