Devstral 2 2512 vs Llama Guard 4 12B

Price, context window & benchmark comparison

Mistral AI

Devstral 2 2512

Mistral AI
$0.40 / $2.00
Input / Output · per 1M tokens
VS
Meta Llama

Llama Guard 4 12B

Meta Llama
$0.18 / $0.18
Input / Output · per 1M tokens
Compare different models
Mistral AI
Meta Llama
Verdict: Llama Guard 4 12B is 4.4× cheaper than Devstral 2 2512 on a typical 3:1 input/output mix. Devstral 2 2512 has the larger context window (262K vs 164K).
Devstral 2 2512Llama Guard 4 12B
Input price / 1M$0.40$0.18
Output price / 1M$2.00$0.18
Blended price (3:1)$0.80$0.18
Cached input / 1M$0.04—
Context window262K164K
Max output210K16K
Intelligence index——
Coding index31.3—
Providers13
Best provider uptime (30m)100.0%100.0%
Vision inputNoYes
ReasoningNoNo
Tool callingYesNo
Released2025-12-092025-04-30
Retirement——

Cost calculator

FAQ

Which is cheaper, Devstral 2 2512 or Llama Guard 4 12B?

Llama Guard 4 12B is 4.4× cheaper than Devstral 2 2512 on a typical 3:1 input/output mix.

Devstral 2 2512 vs Llama Guard 4 12B: which has a bigger context window?

Devstral 2 2512: 262,144 tokens. Llama Guard 4 12B: 163,840 tokens.

How much would 1 million requests cost?

At 2,000 input + 500 output tokens per request: Devstral 2 2512 ≈ $1,800, Llama Guard 4 12B ≈ $450.00.

More comparisons