Mara — pricing & performance
Models served
6
of 1,128 tracked models
Mean output speed
111
tokens per second, averaged over
5 measured endpoints
5 measured endpoints
Mean latency
2.5s
time to first token, averaged over
5 measured endpoints
5 measured endpoints
Cheapest blended
$0.30
per 1M tokens at a 3:1
input:output blend
input:output blend
Output speed on Mara
Output tokens per second for the 5 of 6 models on Mara with published throughput telemetry, the most recent figure OpenRouter publishes for each host. The full table below lists all 6, measured or not. Higher is better. We do not run these measurements ourselves.
All 5 of 5 models⇩
The Known Good
6 models on Mara
Full leaderboard →
| Model | Creator | Input $/M | Output $/M | Blended | Speed | TTFT | Context |
|---|---|---|---|---|---|---|---|
DeepSeek: DeepSeek V3.1
|
DeepSeek | $0.60 | $1.70 | $0.88 | 50 | 2.3s | 131k |
|
DeepSeek: DeepSeek V3.2
|
DeepSeek | $3.00 | $4.50 | $3.38 | — | — | 33k |
MiniMax: MiniMax M2.5
|
MiniMax | $0.24 | $0.96 | $0.42 | 227 | 1.0s | 197k |
|
MiniMax: MiniMax M2.7
|
MiniMax | $0.24 | $0.96 | $0.42 | 113 | 6.4s | 197k |
|
MiniMax: MiniMax M3
|
MiniMax | $0.60 | $2.40 | $1.05 | 48 | 1.7s | 1.05M |
OpenAI: gpt-oss-120b
|
OpenAI | $0.15 | $0.75 | $0.30 | 119 | 0.9s | 131k |
The Known Good
Source. Prices, endpoint availability and throughput for Mara 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.