Providers/Together
Together logo

Together

Inference providertogether.aiStatus ↗
MODELS
21
ENDPOINTS
22
MIN INPUT
$0.05
MEDIAN BLENDED
$0.656
AVG UPTIME
93.4%
CHEAPEST FOR
2 models
Models on Together · prices & measurements for models served by Together’s APIUSD / MTOK · VERIFIED 4 MINUTES AGO
MODELINPUTOUTPUTCONTEXTQUANTTTFTTPSSMOKEVS BEST ELSEWHERE
OpenAI logogpt-oss-20bOpenAI$0.05$0.20131K$0.03 elsewhere
Google logoGemma 3n 4BGoogle$0.06$0.1233K1.3s°53° tok/sonly provider
DeepSeek logoDeepSeek V4 Flash 0731DeepSeek$0.14$0.281.05M0.68s°137° tok/s$0.04 elsewhere
OpenAI logogpt-oss-120bOpenAI$0.15$0.60131K0.54s°88° tok/s$0.03 elsewhere
Qwen logoQwen3.5-9BQwen$0.17$0.25262K0.86s°95° tok/s$0.10 elsewhere
Meta logoLlama Guard 4 12BMeta$0.20$0.201.05M0.76s°$0.18 elsewhere
Google logoGemma 4 31BGoogle$0.28$0.86262K2.2s°25° tok/s$0.08 elsewhere
MiniMax logoMiniMax M3MiniMax$0.30$1.20524K0.37s°188° tok/s$0.23 elsewhere
Qwen logoQwen2.5 7B InstructQwen$0.30$0.3033K$0.10 elsewhere
Meta logoMuse Glimmer 30BMeta$0.35$1.50131K6.2s°51° tok/s$0.30 elsewhere
TMInkling SmallThinking Machines$0.50$1.20524K0.51s°88° tok/s$0.45 elsewhere
NVIDIA logoNemotron 3 UltraNVIDIA$0.60$3.60512K0.64s°153° tok/s$0.50 elsewhere
AAVirtuoso LargeArcee AI$0.75$1.20131Konly provider
Moonshot AI logoKimi K2.7 CodeMoonshot AI$0.95$4.00262K0.39s°114° tok/s$0.67 elsewhere
TMInklingThinking Machines$1.00$4.05524K0.68s°151° tok/s$0.95 elsewhere
Meta logoLlama 3.3 70B InstructMeta$1.04$1.04131K1.6s°58° tok/s$0.10 elsewhere
DeepSeek logoDeepSeek V4 Pro 0813DeepSeek$1.32$3.961.05M0.76s°115° tok/s4/5$0.66 elsewhere
Z.ai logoGLM 5.2Z.ai$1.40$4.40512K0.54s°131° tok/s$0.50 elsewhere
DeepSeek logoDeepSeek V4 Pro 0423DeepSeek$1.74$3.48512K1.1s°120° tok/s$0.522 elsewhere
Qwen logoQwen3.8 2.4T A95BQwen$2.50$6.251.01M1.2s°191° tok/s4/5$2.00 elsewhere
Moonshot AI logoKimi K3Moonshot AI$3.00$15.001.05M5s°24° tok/s$2.60 elsewhere
TTFT/TPS: ModelIndex-measured medians (P50, trailing 72h) for Together’s endpoints · ° = measured via OpenRouter routing · ✕ = probed, no measurable stream · SMOKE: of 5 capability probes passed · methodology on each model page