Benchmarks / Artificial Analysis

Reported by Artificial Analysis

Artificial Analysis

Median tokens a second while the model writes its answer, measured by Artificial Analysis on the model's usual API over the past weeks.

Last updated 8 Oct 2026

Results dated
8 Oct 2026
Results
133 configurations of 74 models
Unit
tokens per second
Licence
Artificial Analysis commercial data licence

Output speed: Qwen3.5-Omni-Plus

Top 15 of 74 results · tokens per second, higher is better. Choose a model to highlight it.Clear highlight

  1. 1 Trinity Large ThinkingArcee AI 347
  2. 2 Gemini 3.5 Flash-LiteGoogle 335
  3. 3 Nova Micro 1.0Amazon 262
  4. 4 gpt-oss-20b (low reasoning)OpenAI 255
  5. 5 Claude Haiku 5.5 (max reasoning)Anthropic 235
  6. 6 Ministral 3 3BMistral AI 233
  7. 7 Qwen3.5-Omni-FlashAlibaba 226
  8. 8 Nemotron 3 Nano 30B A3B (no reasoning)NVIDIA 223
  9. 9 DeepSeek-V4.1-Flash (max reasoning)DeepSeek 217
  10. 9 Nova 2 Lite (no reasoning)Amazon 217
  11. 11 DeepSeek-V4-Flash-Vision-Exp (max reasoning)DeepSeek 215
  12. 12 gpt-oss-120b (low reasoning)OpenAI 188
  13. 13 Qwen3-Next-80B-A3B-InstructAlibaba 179
  14. 14 Qwen3-Next-80B-A3B-ThinkingAlibaba 178
  15. 15 Inkling SmallThinking Machines 173
  16. 35 Qwen3.5-Omni-PlusAlibaba 101

Full results

Artificial Analysis: Output speed, tokens per second, higher is better
#ModelOutput speed
tokens per second, higher is better		Price
$ per million tokens, in / out
1 Trinity Large ThinkingArcee AI
347
$0.25 / $0.80
2 Gemini 3.5 Flash-LiteGoogle
335
$0.30 / $2.50
3 Nova Micro 1.0Amazon
262
$0.035 / $0.14
4 gpt-oss-20b (low reasoning)OpenAI · best of 2 settings
255
$0.03 / $0.15
5 Claude Haiku 5.5 (max reasoning)Anthropic · best of 5 settings
235
$0.10 / $0.50
6 Ministral 3 3BMistral AI
233
$0.10 / $0.10
7 Qwen3.5-Omni-FlashAlibaba
226
–
8 Nemotron 3 Nano 30B A3B (no reasoning)NVIDIA · best of 2 settings
223
$0.05 / $0.20
9 DeepSeek-V4.1-Flash (max reasoning)DeepSeek · best of 2 settings
217
$0.30 / $1.20
9 Nova 2 Lite (no reasoning)Amazon · best of 4 settings
217
$0.30 / $2.50
11 DeepSeek-V4-Flash-Vision-Exp (max reasoning)DeepSeek
215
$0.44 / $1.32
12 gpt-oss-120b (low reasoning)OpenAI · best of 2 settings
188
$0.15 / $0.60
13 Qwen3-Next-80B-A3B-InstructAlibaba
179
$0.10 / $1.10
14 Qwen3-Next-80B-A3B-ThinkingAlibaba
178
$0.15 / $1.20
15 Inkling SmallThinking Machines
173
$0.45 / $1.20
16 Mistral Small 4Mistral AI · best of 2 settings
164
$0.15 / $0.60
17 Mistral Medium 3.5Mistral AI
160
$1.50 / $7.50
18 Inkling (extra-high reasoning)Thinking Machines
153
$0.95 / $4.05
19 Nemotron 3 Super 120B A12BNVIDIA
148
$0.085 / $0.40
20 Gemini 3.8 Flash (high reasoning)Google
147
$1.50 / $7.50
20 Qwen3.5-122B-A10B (no reasoning)Alibaba · best of 2 settings
147
$0.26 / $2.08
22 Gemma 4 12BGoogle · best of 2 settings
146
–
23 Qwen3.6-35B-A3B (no reasoning)Alibaba · best of 2 settings
138
$0.10 / $1
24 Claude Sonnet 5.5 (max reasoning)Anthropic · best of 5 settings
134
$2 / $10
25 GPT-6 Luna (max reasoning)OpenAI · best of 5 settings
127
$0.10 / $0.50
26 o3OpenAI
123
$2 / $8
27 Gemini 3.1 Pro PreviewGoogle
120
$2 / $12
28 GPT-5.5 Instant (2026-06-26)OpenAI
116
–
29 Muse Spark 1.3 (extra-high reasoning)Meta · best of 2 settings
114
$1.25 / $4.25
30 Qwen3-Omni-30B-A3B-ThinkingAlibaba
112
–
31 Qwen3-Omni-30B-A3B-InstructAlibaba
108
–
32 Mistral Large 4Mistral AI
106
$1.36 / $4.18
32 DeepSeek-V4-Pro (0813, no reasoning)DeepSeek · best of 2 settings
106
$1.32 / $3.96
34 Ministral 3 8BMistral AI
102
$0.15 / $0.15
35 Qwen3.5-Omni-PlusAlibaba
101
–
36 GPT-5.6 Terra (max reasoning)OpenAI · best of 6 settings
99.1
$2 / $12
37 Claude Opus 5.5 (max reasoning)Anthropic · best of 5 settings
97.4
$4 / $20
38 Claude Haiku 4.5 (reasoning on)Anthropic · best of 2 settings
97.2
$1 / $5
39 Reka Flash 3Reka AI
94.9
$0.10 / $0.20
40 Qwen3-Coder-NextAlibaba
94.1
$0.18 / $0.90
41 Gemma 4 26B A4B (no reasoning)Google
90.7
$0.10 / $0.30
42 Qwen3.5-397B-A17BAlibaba · best of 2 settings
90.0
$0.55 / $3.50
43 GPT-5.3-Codex (extra-high reasoning)OpenAI
89.2
$1.75 / $14
44 Hy3Tencent
87.5
$0.14 / $0.58
45 Llama 3.3 70B InstructMeta
84.6
$0.59 / $0.79
46 Qwen3.5-9B (no reasoning)Alibaba · best of 2 settings
84.3
$0.10 / $0.15
47 Kimi K2.7 CodeMoonshot AI
83.2
$0.95 / $4
48 MiniMax-M3MiniMax
82.0
$0.30 / $1.20
49 Solar Pro 4Upstage
81.6
$0.09 / $0.36
50 Ministral 3 14BMistral AI
81.3
$0.20 / $0.20
51 Muse Glimmer (high reasoning)Meta
80.4
–
52 Mistral Large 3Mistral AI
78.0
$0.50 / $1.50
53 GLM-5.3 (low reasoning)Z.ai · best of 2 settings
76.3
$1.40 / $4.40
54 Claude Fable 5.1 (max reasoning)Anthropic · best of 5 settings
67.1
$10 / $50
55 Grok 4.7 (high reasoning)xAI · best of 3 settings
63.7
$2 / $6
56 Command ACohere
59.7
$2.50 / $10
57 Qwen3.7-PlusAlibaba
55.4
$0.32 / $1.28
58 GPT-6.1 Sol (max reasoning)OpenAI · best of 5 settings
55.2
$2 / $10
59 Qwen3.8-Flash-NextAlibaba
54.9
–
60 MiMo-V2.6-FlashXiaomi
52.4
$0.14 / $0.28
61 Qwen3.8-27B (low reasoning)Alibaba · best of 4 settings
51.9
$0.50 / $3
62 GPT-6 Astra (extra-high reasoning)OpenAI · best of 5 settings
51.7
$10 / $50
63 Granite 4.2 8BIBM
51.6
$0.06 / $0.25
64 GLM-5.3-FlashZ.ai
49.5
$0.15 / $0.50
65 Llama 4 ScoutMeta
45.6
$0.18 / $0.59
66 Kimi K3 (max reasoning)Moonshot AI · best of 2 settings
45.0
$3 / $15
67 Phi-4Microsoft
43.7
$0.07 / $0.14
68 Gemma 4 31B (no reasoning)Google · best of 2 settings
42.9
$0.14 / $0.40
69 MiMo-V2.6-ProXiaomi
39.0
$0.43 / $0.87
70 Qwen3.8-2.4T-A95BAlibaba
37.5
$2 / $6
71 Qwen3.8-Max (0902)Alibaba
37.0
$2 / $6
72 Llama 4 MaverickMeta
32.8
$0.27 / $0.85
73 Gemma 4 E4BGoogle · best of 2 settings
27.3
–
74 Qwen3.5-4B (reasoning on)Alibaba · best of 2 settings
23.9
–

Swipe the table sideways for more columns.

Ranks follow the score as shown, so equal numbers share a rank. Each model is shown at its best setting; show every setting. Results as published by Artificial Analysis; we do not re-run them.

What it measures

Median tokens a second while the model writes its answer, measured by Artificial Analysis on the model's usual API over the past weeks.

What it does not measure

Not time spent thinking before the answer starts, or speed on other hosts.

Source

Artificial Analysis · Methodology

282 results from Artificial Analysis not ranked here · show why

We rank a result only when we can tie it to a specific model you can use. These are left out:

  • Not on sale through the API providers we track: 268
  • A different snapshot or variant from the model we list: 13
  • An unusual combination of settings: 1

Data sourced from Artificial Analysis. Licence: Artificial Analysis commercial data licence.