The Known Good Updated 29 Jul 2026
Providers· First-party endpoint· 17 tracked models

Mistral — pricing & performance

Compare models All providers Data sources
Models served
17
of 947 tracked models
Mean output speed
54
tokens per second, averaged over
17 measured endpoints
Mean latency
0.4s
time to first token, averaged over
17 measured endpoints
Cheapest blended
$0.10
per 1M tokens at a 3:1
input:output blend
Output speed on Mistral

Median output tokens per second for all 17 models on Mistral with published throughput telemetry from OpenRouter. Higher is better. We do not run these measurements ourselves.

⇩
131
120
98
76
52
48
47
45
45
43
42
42
The Known Good
17 models on Mistral
Full leaderboard →
Model Creator Input $/M Output $/M Blended Speed TTFT Context
Mistral Large Mistral $2.00 $6.00 $3.00 33 0.5s 128k
Mistral Large 2407 Mistral $2.00 $6.00 $3.00 24 0.6s 131k
Mistral: Codestral 2508 Mistral $0.30 $0.90 $0.45 43 0.3s 256k
Mistral: Devstral 2 2512 Mistral $0.40 $2.00 $0.80 7 0.4s 262k
Mistral: Ministral 3 14B 2512 Mistral $0.20 $0.20 $0.20 52 0.3s 262k
Mistral: Ministral 3 3B 2512 Mistral $0.10 $0.10 $0.10 47 0.3s 131k
Mistral: Ministral 3 8B 2512 Mistral $0.15 $0.15 $0.15 42 0.3s 262k
Mistral: Mistral Large 3 2512 Mistral $0.50 $1.50 $0.75 45 0.5s 262k
Mistral: Mistral Medium 3 Mistral $0.40 $2.00 $0.80 42 0.5s 131k
Mistral: Mistral Medium 3.1 Mistral $0.40 $2.00 $0.80 45 0.3s 131k
Mistral: Mistral Medium 3.5 Mistral $1.50 $7.50 $3.00 30 0.4s 262k
Mistral: Mistral Nemo Mistral $0.15 $0.15 $0.15 48 0.3s 131k
Mistral: Mistral Small 3.2 24B Mistral $0.10 $0.30 $0.15 98 0.3s 131k
Mistral: Mistral Small 4 Mistral $0.15 $0.60 $0.26 120 0.4s 262k
Mistral: Mixtral 8x22B Instruct Mistral $2.00 $6.00 $3.00 76 0.5s 66k
Mistral: Saba Mistral $0.20 $0.60 $0.30 42 0.4s 33k
Mistral: Voxtral Small 24B 2507 Mistral $0.10 $0.30 $0.15 131 0.4s 32k
The Known Good
Source. Prices, endpoint availability and throughput for Mistral 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.