Friendli — pricing & performance
Models served
6
of 947 tracked models
Mean output speed
65
tokens per second, averaged over
6 measured endpoints
6 measured endpoints
Mean latency
0.7s
time to first token, averaged over
6 measured endpoints
6 measured endpoints
Cheapest blended
$0.20
per 1M tokens at a 3:1
input:output blend
input:output blend
Output speed on Friendli
Median output tokens per second for all 6 models on Friendli with published throughput telemetry from OpenRouter. Higher is better. We do not run these measurements ourselves.
All 6 of 6 models⇩
The Known Good
6 models on Friendli
Full leaderboard →
| Model | Creator | Input $/M | Output $/M | Blended | Speed | TTFT | Context |
|---|---|---|---|---|---|---|---|
DeepSeek: DeepSeek V3.2
|
DeepSeek | $0.50 | $1.50 | $0.75 | 40 | 0.4s | 164k |
Google: Gemma 4 31B
|
$0.14 | $0.40 | $0.20 | 70 | 1.8s | 262k | |
MiniMax: MiniMax M2.5
|
MiniMax | $0.30 | $1.20 | $0.53 | 90 | 0.6s | 197k |
Qwen: Qwen3 235B A22B Instruct 2507
|
Alibaba | $0.20 | $0.80 | $0.35 | 30 | 0.3s | 262k |
Z.ai: GLM 5.1
|
Z.ai | $1.40 | $4.40 | $2.15 | 59 | 0.4s | 203k |
|
Z.ai: GLM 5.2
|
Z.ai | $1.40 | $4.40 | $2.15 | 101 | 0.9s | 1.05M |
The Known Good
Source. Prices, endpoint availability and throughput for Friendli 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.