Novita vs Together AI

67 vs 29 models, 17 shared

Shared models

Model	Novita tok/s	Together AI tok/s
Qwen3-235B-A22B-Instruct-2507	34 tok/s	23 tok/s
Qwen3-Coder-480B-A35B-Instruct	64 tok/s	58 tok/s
Qwen3-Next-80B-A3B-Instruct	101 tok/s	136 tok/s
Qwen3-VL-8B-Instruct	67 tok/s	63 tok/s
Qwen3.5-397B-A17B	47 tok/s	18 tok/s
DeepSeek-R1	40 tok/s	68 tok/s
DeepSeek-R1-0528	38 tok/s	84 tok/s
DeepSeek-V3	33 tok/s	42 tok/s
DeepSeek-V3-0324	32 tok/s	31 tok/s
DeepSeek-V3.1	36 tok/s	40 tok/s
Llama-3.3-70B-Instruct	27 tok/s	108 tok/s
Llama-4-Maverick-17B-128E-Instruct-FP8	89 tok/s	48 tok/s
Kimi-K2.5	102 tok/s	64 tok/s
gpt-oss-120b	51 tok/s	81 tok/s
gpt-oss-20b	97 tok/s	76 tok/s
GLM-4.6	100 tok/s	42 tok/s
GLM-5	39 tok/s	42 tok/s