Devstral Pricing Plans

ComponentCostDetails
Devstral (open weights)FreeDownload and run locally
Mistral API (small)$0.10/M input, $0.30/M outputBudget option via Mistral platform
Mistral API (large)$2.00/M input, $6.00/M outputPremium option

Best AI Models for Devstral (2026)

Based on cost-effectiveness, coding quality, and availability. Data updated Apr 18, 2026.

💰 Cheapest

Mistral Small 3

Mistral · $0.100/M input

High-volume tasks, cost optimization

View full pricing →
⚡ Best Value

Mistral Nemo

Mistral · $0.150/M input

Self-hosted deployments, cost-sensitive coding, edge deployments

View full pricing →
🏆 Premium

Mistral Large 3

Mistral · $2.00/M input

Enterprise coding, European data sovereignty, multilingual tasks

View full pricing →

Cost Per Session with Devstral

Estimated cost for a medium feature implementation (100K input + 10K output tokens).

ModelProviderCost/SessionInput PriceOutput Price
Mistral Nemo Mistral $0.08 $0.150/M $0.150/M
Mistral Small 3 Mistral $0.10 $0.100/M $0.300/M
Codestral Mistral $0.29 $0.300/M $0.900/M
Mistral Medium Mistral $0.54 $0.400/M $2.00/M
Mistral Large 2 Mistral $1.90 $2.00/M $6.00/M
Mistral Large 3 Mistral $1.90 $2.00/M $6.00/M

Frequently Asked Questions

Is Devstral free?

Yes, Devstral is open-weight. You can download and run it locally for free, or use the Mistral API at competitive rates.

How much does Devstral cost via API?

Via Mistral's API, coding sessions cost as little as $0.01 with the small model, or $0.15-$0.30 with larger models.

Can I self-host Devstral?

Yes, the weights are publicly available. You can run it on your own hardware via Ollama, vLLM, or other inference engines.

Compare Devstral with Other Platforms

See how Devstral stacks up against the competition.

View all AI coding platforms →