modelgrep

Anthropic: Claude 3 Haiku vs Meta: Llama 3.1 70B Instruct

Meta: Llama 3.1 70B Instruct wins on more metrics (5 of 8), but the right pick depends on what you optimize for — see the breakdown below.

MetricAnthropic: Claude 3 HaikuMeta: Llama 3.1 70B Instruct
Intelligence Index12.312.5
Coding Index6.710.9
GPQA Diamond37%41%
Design Arena Elo
Speed (tokens/sec)7929
Latency481ms293ms
Input price /M$0.250$0.400
Output price /M$1.25$0.400
Context window200K131K
CapabilitiesToolsVisionToolsJSON