Speed

How fast output arrives — sustained throughput and time to the first token, which trade off against each other and against cost. Measured on a provider's serving of the model, so it describes an endpoint rather than the weights.

Ordered by how many of these benchmarks put a model in their top 10, then by its best placing. Each column keeps its own ranking; nothing is averaged.

Model	In top 10	Artificial Analysis LLM Leaderboard Output Speed	Artificial Analysis LLM Leaderboard Time To First Token
Celeris 1 celeris	2/2	#1 1523/s	#9 0.58s
Gemini 2.5 Flash Lite google	2/2	#5 360/s	#1 0.30s
Nemotron 3.5 Lightning nvidia	2/2	#9 305/s	#8 0.56s
Mercury 2.5 inception	1/2	#2 691/s	#134 3.44s
Nemotron 3 Nano Omni nvidia	1/2	#15 271/s	#2 0.35s
Mercury 2 inception	1/2	#3 632/s	#148 7.12s
North Mini Code cohere	1/2	#111 83/s	#3 0.45s
Trinity Large Thinking arcee	1/2	#4 362/s	#61 1.22s
Gemini 2.5 Flash google	1/2	#27 210/s	#4 0.46s
Command A+ cohere	1/2	#57 144/s	#5 0.48s
Gemini 3.5 Flash Lite google	1/2	#6 347/s	#151 9.42s
Granite 4.2 ibm	1/2	#25 216/s	#6 0.53s
Ling 3.0 Flash Fin inclusionai	1/2	#7 335/s	#83 1.75s
Grok Build 0.1 xai	1/2	#116 80/s	#7 0.54s
Ling 3.0 Flash inclusionai	1/2	#8 323/s	#111 2.52s
Muse Spark 1.2 meta	1/2	#10 301/s	#154 15.17s
GPT 5.4 mini openai	1/2	#14 273/s	#10 0.61s

A dash means the board does not list that model — not a score of zero. Entries not yet matched to a model are left out here; they still appear on their own board.