Providers· Inference host· 9 tracked models

Groq — pricing & performance

Compare models All providers Data sources

Models served

of 1,109 tracked models

Mean output speed

207

tokens per second, averaged over
9 measured endpoints

Mean latency

0.3s

time to first token, averaged over
9 measured endpoints

Cheapest blended

$0.058

per 1M tokens at a 3:1
input:output blend

Output speed on Groq

Output tokens per second for all 9 models on Groq with published throughput telemetry, the most recent figure OpenRouter publishes for each host. Higher is better. We do not run these measurements ourselves.

All 9 of 9 models⇩

319

315

304

266

208

205

122

gpt-oss-20b

gpt-oss-safeguard-20b

Llama 3.3 70B
Instruct

Llama 3.1 8B
Instruct

Kimi K2 0905

The Known Good

9 models on Groq

Full leaderboard →

Model	Creator	Input $/M	Output $/M	Blended	Speed	TTFT	Context
Meta: Llama 3.1 8B Instruct	Meta	$0.050	$0.080	$0.058	63	0.4s	131k
Meta: Llama 3.3 70B Instruct	Meta	$0.59	$0.79	$0.64	122	0.2s	131k
Meta: Llama 4 Scout	Meta	$0.11	$0.34	$0.17	205	0.3s	131k
MiniMax: MiniMax M2.7	MiniMax	$0.60	$1.80	$0.90	266	0.3s	197k
MoonshotAI: Kimi K2 0905	Moonshot AI	$1.00	$3.00	$1.50	59	0.4s	262k
OpenAI: gpt-oss-120b	OpenAI	$0.15	$0.60	$0.26	304	0.2s	131k
OpenAI: gpt-oss-20b	OpenAI	$0.075	$0.30	$0.13	319	0.5s	131k
OpenAI: gpt-oss-safeguard-20b	OpenAI	$0.075	$0.30	$0.13	315	0.3s	131k
Qwen: Qwen3 32B	Alibaba	$0.29	$0.59	$0.36	208	0.4s	131k

The Known Good

Source. Prices, endpoint availability and throughput for Groq 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.