modelgrep

NVIDIA: Llama 3.3 Nemotron Super 49B V1.5 vs OpenAI: GPT-4o (2024-08-06)

NVIDIA: Llama 3.3 Nemotron Super 49B V1.5 wins on more metrics (5 of 8), but the right pick depends on what you optimize for — see the breakdown below.

MetricNVIDIA: Llama 3.3 Nemotron Super 49B V1.5OpenAI: GPT-4o (2024-08-06)
Intelligence Index14.618.6
Coding Index10.516.6
GPQA Diamond48%52%
Design Arena Elo
Speed (tokens/sec)4130
Latency160ms893ms
Input price /M$0.400$2.50
Output price /M$0.400$10.00
Context window131K128K
CapabilitiesReasoningToolsJSONToolsJSONVision