Provider throughput

Median provider output throughput in tokens per second across OpenRouter endpoints (3d).

182 models · 182 results · data as of 2026-11-06 · openrouter.ai ↗

All results

182 rows

#	Model	Effort	Provider throughput
1	Muse Spark 1.2	max	210.5
2	Muse Spark 1.1	max	192.0
3	Gemini 2.5 Flash Lite	none	174.8
4	Llama 3.2 3b	none	161.5
5	Haiku 5.5	max	146.8
6	Grok Build 0.1		136.4
7	Gemini 3.5 Flash-Lite	xhigh	134.5
8	Qwen 3 30B-A3B	none	128.5
9	Gemini 3.1 Flash-Lite	high	122.0
10	GPT-5.6 Luna Pro		122.0
11	Qwen 2.5 7B		119.0
12	Mistral Medium 3.5	high	118.1
13	Codestral 2508		117.8
14	Gemini 3.6 Flash	xhigh	117.5
15	gpt-oss-120b	high	117.1
16	DeepSeek V4.1 Flash	max	114.8
17	DeepSeek V4.1 Flash 1017	max	112.5
18	gpt-oss-20b	high	110.0
19	Gemini 3.5 Flash	xhigh	109.0
20	Mistral Small 4	high	108.0
21	Mistral Medium 3	none	106.5
22	Muse Spark Open	xhigh	106.0
23	Llama 3.2 1b	none	105.0
24	GPT-5.6 Sol Pro	max	104.0
25	MiniMax M3.1	max	101.5
26	Sonnet 5.5	max	100.0
27	GPT-6 Luna Pro	max	99.8
28	GPT-5.6 Luna	max	99.4
29	Gemini 3 Pro	high	98.5
30	GPT-5.6 Terra Pro		96.0