modelgrep

Meta: Llama 3.3 70B Instruct vs MoonshotAI: Kimi K2 0711

MMeta: Llama 3.3 70B Instructmeta-llama/llama-3.3-70b-instructVSMMoonshotAI: Kimi K2 0711moonshotai/kimi-k2

Meta: Llama 3.3 70B Instruct wins on more metrics (6 of 8), but the right pick depends on what you optimize for — see the breakdown below.

MetricMeta: Llama 3.3 70B InstructMoonshotAI: Kimi K2 0711
Intelligence Index14.5✓14.4
Coding Index10.7✓10.5
GPQA Diamond50%54%✓
Design Arena Elo—1095✓
Speed (tokens/sec)116✓15
Latency198ms✓1.6s
Input price /M$0.100✓$0.570
Output price /M$0.320✓$2.30
Context window131K131K
CapabilitiesToolsJSONTools
Meta: Llama 3.3 70B Instruct details →MoonshotAI: Kimi K2 0711 details →