Skip to main content

Llama 3.2 1B Instruct vs Phi-4

SpecLlama 3.2 1B InstructPhi-4
Respan Index5.3 (#176)9.2 (#174)
Blended price per 1M$0.02$0.088
Context window128K16K
Input / output per 1M$0.02 / $0.02$0.07 / $0.14
Max output-16K
Inputtexttext
Released2024-09-252024-12-12

Respan Index by category

Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.

CategoryLlama 3.2 1B InstructPhi-4
Coding0.0* Llama 3.2 1B Instruct0.6* Phi-4
Agents0.2 Llama 3.2 1B Instruct0.9 Phi-4
Reasoning12.5 Llama 3.2 1B Instruct17.2 Phi-4
Knowledge & language0.3* Llama 3.2 1B Instruct4.9 Phi-4
Human preference13.8 Llama 3.2 1B Instruct27.1 Phi-4

* Estimated: no published scores in that category. How the Respan Index works

Benchmarks both report

Phi-4 leads on 6 of the 6 benchmarks both report. Higher is better. Each score links to where it was published.

BenchmarkLlama 3.2 1B InstructPhi-4
LMArena Elo11111256
GPQA Diamond23.9%56.1%
OTIS Mock AIME 2024-2025 (Epoch AI)0.6%13.8%
BALROG (BALROG)6.6%11.6%
IFEval59.5%63%
MMLU49.3%84.8%

FAQ

Is Llama 3.2 1B Instruct better than Phi-4?

Phi-4 scores 9.2 on the Respan Index (#174), ahead of Llama 3.2 1B Instruct at 5.3 (#176). Phi-4 leads on 6 of the 6 benchmarks both report.

Which is cheaper, Llama 3.2 1B Instruct or Phi-4?

Llama 3.2 1B Instruct lists at $0.02 input and $0.02 output per 1M tokens; Phi-4 at $0.07 and $0.14. Llama 3.2 1B Instruct is 4.4x cheaper per token at list price.

Which has the larger context window?

Llama 3.2 1B Instruct accepts 128K tokens and Phi-4 16K tokens.

Compare other models

vs

Built for AI agents.
Break less.
Ship more.