← All comparisons

Llama 3.1 Instruct 405B vs Llama 3.3 Nemotron Super 49B v1 (Reasoning)

Meta vs NVIDIA — side-by-side benchmark comparison

Llama 3.1 Instruct 405BLlama 3.3 Nemotron Super 49B v1 (Reasoning)
Intelligence Index17.418.5
Coding Index14.59.4
Math Index3.054.7
Output speed (tok/s)37.50.0
Blended price ($/1M)$3.69$0.00
Time to first token (s)0.63s0.00s
aime21.3%58.3%
aime 253.0%54.7%
artificial analysis coding index14.509.40
artificial analysis intelligence index17.4018.50
artificial analysis math index3.0054.70
gpqa51.5%64.3%
hle4.2%6.5%
ifbench39.0%38.1%
lcr24.3%17.0%
livecodebench30.5%27.7%
math 50070.3%95.9%
mmlu pro73.2%78.5%
scicode29.9%28.2%
tau219.0%26.9%
terminalbench hard6.8%0.0%

Benchmark data from Artificial Analysis.