The Known Good Updated 29 Jul 2026
Providers· Inference host· 14 tracked models

Cloudflare — pricing & performance

Compare models All providers Data sources
Models served
14
of 947 tracked models
Mean output speed
50
tokens per second, averaged over
14 measured endpoints
Mean latency
0.7s
time to first token, averaged over
14 measured endpoints
Cheapest blended
$0.041
per 1M tokens at a 3:1
input:output blend
Output speed on Cloudflare

Median output tokens per second for all 14 models on Cloudflare with published throughput telemetry from OpenRouter. Higher is better. We do not run these measurements ourselves.

⇩
137
74
65
64
60
56
53
38
34
33
28
24
The Known Good
14 models on Cloudflare
Full leaderboard →
Model Creator Input $/M Output $/M Blended Speed TTFT Context
DeepSeek: DeepSeek V4 Flash DeepSeek $0.14 $0.28 $0.17 34 1.0s 384k
DeepSeek: DeepSeek V4 Pro DeepSeek $1.74 $3.48 $2.17 64 0.7s 393k
Google: Gemma 4 26B A4B Google $0.10 $0.30 $0.15 60 0.5s 256k
IBM: Granite 4.0 Micro IBM $0.017 $0.11 $0.041 28 0.6s 131k
Meta: Llama 3.1 8B Instruct Meta $0.15 $0.29 $0.19 18 0.6s 32k
Meta: Llama 3.2 1B Instruct Meta $0.027 $0.20 $0.070 137 0.2s 60k
Meta: Llama 3.2 3B Instruct Meta $0.051 $0.34 $0.12 65 0.2s 80k
Meta: Llama 3.3 70B Instruct Meta $0.29 $2.25 $0.78 24 0.7s 24k
Mistral: Mistral Small 3.1 24B Mistral $0.35 $0.56 $0.40 38 0.3s 128k
MoonshotAI: Kimi K2.6 Moonshot AI $0.95 $4.00 $1.71 56 0.7s 262k
MoonshotAI: Kimi K2.7 Code Moonshot AI $0.95 $4.00 $1.71 74 0.6s 262k
Qwen2.5 Coder 32B Instruct Alibaba $0.66 $1.00 $0.74 33 0.5s 33k
Z.ai: GLM 4.7 Flash Z.ai $0.060 $0.40 $0.15 20 0.3s 131k
Z.ai: GLM 5.2 Z.ai $1.40 $4.40 $2.15 53 2.4s 262k
The Known Good
Source. Prices, endpoint availability and throughput for Cloudflare 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.