Inference Chip Index

Leaderboard

Each slice is an exact Closed-division comparison. Rankings never mix release, division, workload, scenario, metric, or unit. Only ranking-eligible accepted rows.

WorkloadScenarioUnitAcceptedSlice ID
llama3.1-8bServerTokens/s23v6.0|closed|llama3.1-8b|Server|Tokens/s|official-system
llama3.1-8bInteractiveTokens/s10v6.0|closed|llama3.1-8b|Interactive|Tokens/s|official-system
llama3.1-8bOfflineTokens/s23v6.0|closed|llama3.1-8b|Offline|Tokens/s|official-system
gpt-oss-120bServerTokens/s28v6.0|closed|gpt-oss-120b|Server|Tokens/s|official-system
gpt-oss-120bInteractiveTokens/s4v6.0|closed|gpt-oss-120b|Interactive|Tokens/s|official-system
gpt-oss-120bOfflineTokens/s28v6.0|closed|gpt-oss-120b|Offline|Tokens/s|official-system
deepseek-r1ServerTokens/s18v6.0|closed|deepseek-r1|Server|Tokens/s|official-system
deepseek-r1InteractiveTokens/s3v6.0|closed|deepseek-r1|Interactive|Tokens/s|official-system
deepseek-r1OfflineTokens/s18v6.0|closed|deepseek-r1|Offline|Tokens/s|official-system