Noometry

Model comparison

Llama 3.1-70B vs Phi-4

Phi-4 is the stronger model overall, scoring 30.9 to 29.4 on the Noometry Index.

Last verified . 25 shared benchmarks.

Llama 3.1-70B Meta

29.4

Rank #299 Confirmed

Phi-4 Microsoft

30.9

Rank #269 Confirmed

Summary

  • They share 25 benchmarks with published results for both. Llama 3.1-70B scores higher in 5 categories and Phi-4 in 4 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where Phi-4 leads 32.5 to 24.1.
  • The biggest single-benchmark swing is MATH Level 5: 36.7% for Llama 3.1-70B and 64.9% for Phi-4.
  • Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.40 / $0.40 for Llama 3.1-70B.

Side by side

Llama 3.1-70BPhi-4
ProviderMetaMicrosoft
Noometry Index29.430.9
Released2024-07-232024-12-11
WeightsOpenOpen
Context window128K128K
Max output4K4K
Input $ / M tokens$0.40$0.07
Output $ / M tokens$0.40$0.14
Results tracked3537

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Phi-4 leads

Llama 3.1-70B: 29.8 (#292), Phi-4: 33.8 (#237)

BenchmarkLlama 3.1-70BPhi-4
BigCodeBench Instruct46.1%45.5%
LMArena Coding12601231
BigCodeBench Complete54.8%55.4%
WeirdML9%—
LiveBench Coding—30.7%

Agentic & Tool Use Llama 3.1-70B leads

Llama 3.1-70B: 24.9 (#104), Phi-4: 22.6 (#120)

BenchmarkLlama 3.1-70BPhi-4
BALROG27.9%11.6%
Berkeley Function Calling Leaderboard—28.8%
TheAgentCompany6.9%—

Reasoning Llama 3.1-70B leads

Llama 3.1-70B: 20.8 (#215), Phi-4: 17.0 (#287)

BenchmarkLlama 3.1-70BPhi-4
LMArena Hard Prompts12411220
Epoch Capabilities Index125.91130.43
Chess Puzzles—1%
LiveBench Reasoning—47.8%
DTBench60%—
LiveBench Data Analysis—45.2%
LMCA14.8%—
LiveBench—41.6%

Math Phi-4 leads

Llama 3.1-70B: 12.7 (#298), Phi-4: 19.6 (#279)

BenchmarkLlama 3.1-70BPhi-4
OTIS Mock AIME 2024-20253.6%13.8%
LMArena Math12521246
MATH Level 536.7%64.9%
Omni-MATH21%—
LiveBench Math—42%

Knowledge Phi-4 leads

Llama 3.1-70B: 24.1 (#264), Phi-4: 32.5 (#204)

BenchmarkLlama 3.1-70BPhi-4
GPQA Diamond44.2%56.1%
LMArena Expert12091203
MMLU80.1%84.8%
MMLU-Pro65.3%—
Confabulations—29.4%
Vectara Hallucination Rate—3.7%
GPQA (HELM)42.6%—

Multilingual Llama 3.1-70B leads

Llama 3.1-70B: 38.9 (#220), Phi-4: 37.4 (#232)

BenchmarkLlama 3.1-70BPhi-4
LMArena Non-English12191197
LMArena Chinese12151211
LMArena French12601223
LMArena German12221222
LMArena Japanese11321158
LMArena Korean11401151
LMArena Russian12341209
LMArena Spanish12531234

Instruction Following Llama 3.1-70B leads

Llama 3.1-70B: 65.5 (#218), Phi-4: 60.6 (#246)

BenchmarkLlama 3.1-70BPhi-4
LMArena Instruction Following12311201
LiveBench Instruction Following—58.4%
IFEval82.1%—

Long Context Too close to call

Llama 3.1-70B: 37.6 (#208), Phi-4: 36.9 (#221)

BenchmarkLlama 3.1-70BPhi-4
LMArena Longer Query12411217

Writing & Preference Phi-4 leads

Llama 3.1-70B: 35.6 (#262), Phi-4: 40.7 (#239)

BenchmarkLlama 3.1-70BPhi-4
LMArena Text12611217
LMArena Creative Writing12321182
LMArena Multi-Turn12561206
Short-Story Creative Writing—62.6%
EQ-Bench Creative Writing784—
WildBench75.8%—
LiveBench Language—25.6%

Frequently asked questions

Is Llama 3.1-70B better than Phi-4?

Phi-4 is the stronger model overall, scoring 30.9 to 29.4 on the Noometry Index.

Which is cheaper, Llama 3.1-70B or Phi-4?

Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Llama 3.1-70B lists at $0.40 and $0.40.

Is Llama 3.1-70B or Phi-4 better for coding?

Phi-4 scores higher on coding benchmarks: 33.8 versus 29.8 in the Noometry coding category.

Which has the bigger context window?

Both accept 128K tokens.

How many benchmarks do Llama 3.1-70B and Phi-4 share?

25 benchmarks have published results for both models. Llama 3.1-70B has 35 scored results on Noometry and Phi-4 has 37.

Related comparisons

Go deeper