The Known Good Updated 4 Sep 2026 Subscribe
Providers· Inference host· 13 tracked models

BaseTen — pricing & performance

Compare models All providers Data sources
Models served
13
of 1,094 tracked models
Mean output speed
89
tokens per second, averaged over
12 measured endpoints
Mean latency
1.5s
time to first token, averaged over
12 measured endpoints
Cheapest blended
$0.16
per 1M tokens at a 3:1
input:output blend
Output speed on BaseTen

Output tokens per second for the 12 of 13 models on BaseTen with published throughput telemetry, the most recent figure OpenRouter publishes for each host. The full table below lists all 13, measured or not. Higher is better. We do not run these measurements ourselves.

⇩
187
139
120
96
93
88
80
75
66
46
39
39
The Known Good
13 models on BaseTen
Full leaderboard →
Model Creator Input $/M Output $/M Blended Speed TTFT Context
DeepSeek: DeepSeek V4 Flash 0731 DeepSeek $0.13 $0.26 $0.16 80 0.6s 1.05M
DeepSeek: DeepSeek V4 Pro 0423 DeepSeek $1.74 $3.48 $2.17 66 0.6s 1.05M
DeepSeek: DeepSeek V4 Pro 0813 DeepSeek $1.32 $3.96 $1.98 46 0.5s 1.05M
MoonshotAI: Kimi K2.6 Moonshot AI $0.95 $4.00 $1.71 — — 262k
MoonshotAI: Kimi K2.7 Code Moonshot AI $0.95 $4.00 $1.71 39 0.6s 262k
MoonshotAI: Kimi K3 Moonshot AI $3.00 $15.00 $6.00 39 2.6s 1.05M
NVIDIA: Nemotron 3 Ultra NVIDIA $0.60 $2.40 $1.05 75 0.7s 203k
OpenAI: gpt-oss-120b OpenAI $0.10 $0.50 $0.20 187 0.2s 128k
Thinking Machines: Inkling Thinking Machines $1.00 $4.05 $1.76 120 2.4s 1.05M
Thinking Machines: Inkling Small Thinking Machines $0.50 $1.20 $0.68 139 6.1s 1.05M
Z.ai: GLM 5.2 Z.ai $1.40 $4.40 $2.15 88 1.6s 1.05M
Z.ai: GLM 5.3 Z.ai $1.40 $4.40 $2.15 96 1.5s 1.05M
Z.ai: GLM 5.3 Flash Z.ai $0.15 $0.50 $0.24 93 0.7s 1.05M
The Known Good
Source. Prices, endpoint availability and throughput for BaseTen are ingested from OpenRouter; model metadata is enriched from models.dev. A blank cell means the figure was not published for that endpoint — it never means zero. Blended price is (3 × input + output) ÷ 4, the same 3:1 blend used everywhere on this site; see methodology. Quality scores are not provider-specific: a model scores the same wherever it is hosted, so quality lives on the model pages.