Migration assistant

Replacing Llama 3.3 70B?

Ranked replacement candidates, each shown with exactly what changes if you switch. Same-provider matches surface first because switching cost is lower; capability regressions are flagged in red.

ME
Migrating from

Llama 3.3 70B

Meta · Llama 3 · Active
Input: $0.1/M Output: $0.32/M Context: 128K
View detail →
#1 RECOMMENDED

Llama 4 Scout

Meta · Llama 4
same provider; 20% cheaper input; 10M context (larger); loses 1 capability
View full detail →
What changes if you switch
Field Llama 3.3 70B → Llama 4 Scout Impact
Input price $0.1/M → $0.08/M Save 20%
Output price $0.32/M → $0.3/M Save 6%
Context window 128K tokens → 10M tokens +7712% larger
Function calling ✓ supported → ✗ not supported LOST — re-evaluate before switching
Lifecycle Active → Active Same status
Other candidates
#2

Llama 4 Maverick

Meta
same provider; 50% more expensive input; 1M context (larger)
Details →
Input price: +50% more expensiveOutput price: +88% more expensiveContext window: +681% larger
#3

Amazon Nova Lite

AWS Bedrock
40% cheaper input; 300K context (larger)
Details →
Input price: Save 40%Output price: Save 25%Context window: +134% larger
#4

Amazon Nova Micro

AWS Bedrock
65% cheaper input; same context window
Details →
Input price: Save 65%Output price: Save 56%
#5

Llama 3 8B Instruct

Meta
same provider; 60% cheaper input; loses 1 capability
Details →
Input price: Save 60%Output price: Save 88%Function calling: LOST — re-evaluate before switching
Methodology

How candidates are ranked

Candidates are ranked by how much of the source model's profile each one preserves — weighing switching cost (same provider or model family surface first), capability parity (what you keep versus what you'd lose), context window, and pricing direction. A candidate that is itself deprecated is pushed down the list; one that's already retired is never recommended.

Ranking is purely algorithmic — no editorial weighting, no paid placement. Every value is pulled from each provider's own documentation; click any model name to see the source-linked detail.