LMSpeed Model Speed Leaderboard
Multi-dimensional rankings based on model speed tests, provider health checks, and standard model benchmarks. Compare providers, endpoints, models, and reliability at a glance.
Compares AI model responsiveness across Artificial Analysis output speed, time to first token, and time to first answer token in one table.
Artificial Analysis standard benchmarks
Fairness Notice
These model leaderboards use imported Artificial Analysis benchmark records and are deduplicated to one canonical LMSpeed model per rank.
| Rank | Model | Output Speed | Time to First Token | Time to First Answer Token | Updated |
|---|---|---|---|---|---|
1 | Qwen3.5qwen3-5 | 19.9 tok/s#80 | 0.45 s#7 | 101.19 s#78 | Sep 5, 2026, 04:45 AM |
2 | Phi 4 Multimodal Instructphi-4-multimodal-instruct | 24.6 tok/s#79 | 0.40 s#5 | 0.40 s#3 | Sep 5, 2026, 04:45 AM |
3 | MiMo-V2.5-Promimo-v2-5-pro | 35.8 tok/s#78 | 3.70 s#62 | 59.51 s#75 | Sep 5, 2026, 04:45 AM |
4 | Kimi K3kimi-k3 | 37.9 tok/s#77 | 3.13 s#60 | 55.92 s#74 | Sep 5, 2026, 04:45 AM |
5 | Qwen3.8 Maxqwen3-8-max | 39.1 tok/s#76 | 1.73 s#50 | 52.89 s#72 | Sep 5, 2026, 04:45 AM |
6 | Qwen3.8 2.4T A95Bqwen3-8-2-4t-a95b | 39.8 tok/s#75 | 1.75 s#51 | 52.02 s#71 | Sep 5, 2026, 04:45 AM |
7 | Phi 4phi-4 | 40.8 tok/s#74 | 1.17 s#35 | 1.17 s#18 | Sep 5, 2026, 04:45 AM |
8 | Kimi K2.7 Codekimi-k2-7-code | 41.5 tok/s#73 | 1.27 s#41 | 54.94 s#73 | Sep 5, 2026, 04:45 AM |
9 | Qwen3.8 27Bqwen3-8-27b | 42.4 tok/s#72 | 1.24 s#40 | 48.39 s#70 | Sep 5, 2026, 04:45 AM |
10 | GLM 5.3 Flashglm-5-3-flash | 45.3 tok/s#71 | 1.27 s#42 | 45.44 s#69 | Sep 5, 2026, 04:45 AM |
11 | Longcat 2.0longcat-2-0 | 46.4 tok/s#70 | 2.18 s#56 | 45.28 s#68 | Sep 5, 2026, 04:45 AM |
12 | Phi 4 Mini Instructphi-4-mini-instruct | 46.6 tok/s#69 | 0.36 s#4 | 0.36 s#2 | Sep 5, 2026, 04:45 AM |
13 | Claude Sonnet 4.6claude-sonnet-4-6 | 47.4 tok/s#68 | 1.21 s#37 | 1.21 s#20 | Sep 5, 2026, 04:45 AM |
14 | Grok 4.5grok-4-5 | 50.2 tok/s#67 | 11.88 s#70 | 11.88 s#36 | Sep 5, 2026, 04:45 AM |
15 | Claude Fable 5.1claude-fable-5-1 | 51.7 tok/s#66 | 9.43 s#68 | 9.43 s#33 | Sep 5, 2026, 04:45 AM |
16 | Claude Opus 5claude-opus-5 | 51.9 tok/s#65 | 36.60 s#75 | 36.60 s#62 | Sep 5, 2026, 04:45 AM |
17 | Qwen3.6 27Bqwen3-6-27b | 52.8 tok/s#64 | 1.45 s#46 | 109.00 s#79 | Sep 5, 2026, 04:45 AM |
18 | Command Acommand-a | 54.9 tok/s#63 | 0.35 s#3 | 0.35 s#1 | Sep 5, 2026, 04:45 AM |
19 | Solar Pro 4solar-pro4 | 55.2 tok/s#62 | 1.86 s#53 | 38.11 s#64 | Sep 5, 2026, 04:45 AM |
20 | Qwen3.7 Plusqwen3-7-plus | 55.8 tok/s#61 | 1.52 s#47 | 37.35 s#63 | Sep 5, 2026, 04:45 AM |
21 | MiMo-V2.5mimo-v2-5 | 60.3 tok/s#60 | 3.17 s#61 | 36.34 s#61 | Sep 5, 2026, 04:45 AM |
22 | Grok 4.6grok-4-6 | 60.4 tok/s#59 | 39.85 s#76 | 39.85 s#66 | Sep 5, 2026, 04:45 AM |
23 | Claude Fable 5claude-fable-5 | 65.9 tok/s#58 | 63.82 s#78 | 63.82 s#76 | Sep 5, 2026, 04:45 AM |
24 | DeepSeek V4 Prodeepseek-v4-pro | 66.5 tok/s#57 | 1.16 s#32 | 31.09 s#59 | Sep 5, 2026, 04:45 AM |
25 | Inklinginkling | 67.2 tok/s#56 | 2.50 s#59 | 32.26 s#60 | Sep 5, 2026, 04:45 AM |
26 | North Mini Codenorth-mini-code | 69.4 tok/s#55 | 0.24 s#2 | 29.06 s#56 | Sep 5, 2026, 04:45 AM |
27 | GLM-5.2glm-5-2 | 71.9 tok/s#54 | 1.34 s#44 | 29.15 s#57 | Sep 5, 2026, 04:45 AM |
28 | GPT-5.6 Solgpt-5-6-sol | 75.6 tok/s#53 | 6.91 s#66 | 6.91 s#28 | Sep 5, 2026, 04:45 AM |
29 | Mistral Large 3mistral-large-3 | 76.6 tok/s#52 | 1.29 s#43 | 1.29 s#22 | Sep 5, 2026, 04:45 AM |
30 | GLM 5.3glm-5-3 | 78.4 tok/s#51 | 1.43 s#45 | 26.95 s#55 | Sep 5, 2026, 04:45 AM |
31 | Claude Sonnet 5claude-sonnet-5 | 80.8 tok/s#50 | 100.65 s#79 | 100.65 s#77 | Sep 5, 2026, 04:45 AM |
32 | MiniMax M3minimax-m3 | 84.6 tok/s#49 | 1.17 s#33 | 24.80 s#54 | Sep 5, 2026, 04:45 AM |
33 | Llama 4 Maverickllama-4-maverick | 85.2 tok/s#48 | 0.58 s#14 | 0.58 s#7 | Sep 5, 2026, 04:45 AM |
34 | GPT-5.6 Terragpt-5-6-terra | 86.2 tok/s#47 | 0.85 s#22 | 0.85 s#13 | Sep 5, 2026, 04:45 AM |
35 | Llama 4 Scoutllama-4-scout | 86.6 tok/s#46 | 0.62 s#16 | 0.62 s#9 | Sep 5, 2026, 04:45 AM |
36 | Claude Haiku 4.5claude-haiku-4-5 | 88.5 tok/s#45 | 0.60 s#15 | 0.60 s#8 | Sep 5, 2026, 04:45 AM |
37 | Reka Flash 3reka-flash-3 | 89.2 tok/s#44 | 0.93 s#24 | 23.34 s#51 | Sep 5, 2026, 04:45 AM |
38 | Qwen3.5 397B A17Bqwen3-5-397b-a17b | 89.7 tok/s#43 | 1.80 s#52 | 1.80 s#23 | Sep 5, 2026, 04:45 AM |
39 | Step 3.7 Flashstep-3-7-flash | 91.4 tok/s#42 | 2.33 s#58 | 24.21 s#52 | Sep 5, 2026, 04:45 AM |
40 | Qwen3.5-9Bqwen3-5-9b | 92.2 tok/s#41 | 0.46 s#8 | 0.46 s#5 | Sep 5, 2026, 04:45 AM |
41 | Ministral 3ministral-3 | 93.2 tok/s#40 | 0.47 s#9 | 0.47 s#6 | Sep 5, 2026, 04:45 AM |
42 | Hy3hy3 | 96.0 tok/s#39 | 1.64 s#48 | 22.47 s#49 | Sep 5, 2026, 04:45 AM |
43 | Inkling Smallinkling-small | 97.8 tok/s#38 | 2.22 s#57 | 22.66 s#50 | Sep 5, 2026, 04:45 AM |
44 | GPT-5.6 Lunagpt-5-6-luna | 110.2 tok/s#37 | 109.24 s#80 | 109.24 s#80 | Sep 5, 2026, 04:45 AM |
45 | Grok 4.3grok-4-3 | 112.1 tok/s#36 | 0.82 s#21 | 0.82 s#12 | Sep 5, 2026, 04:45 AM |
46 | Qwen3 Coder Nextqwen3-coder-next | 112.3 tok/s#35 | 0.93 s#25 | 0.93 s#14 | Sep 5, 2026, 04:45 AM |
47 | Gemini 3.1 Pro Previewgemini-3-1-pro-preview | 117.5 tok/s#34 | 24.69 s#74 | 24.69 s#53 | Sep 5, 2026, 04:45 AM |
48 | Granite 4.1 8Bgranite-4-1-8b | 117.9 tok/s#33 | 0.44 s#6 | 0.44 s#4 | Sep 5, 2026, 04:45 AM |
49 | Qwen3.5 Omni Plusqwen3-5-omni-plus | 121.5 tok/s#32 | 1.21 s#37 | 1.21 s#20 | Sep 5, 2026, 04:45 AM |
50 | O3o3 | 123.3 tok/s#31 | 5.17 s#64 | 5.17 s#25 | Sep 5, 2026, 04:45 AM |
51 | Devstral 2devstral-2 | 124.0 tok/s#30 | 0.68 s#18 | 0.68 s#10 | Sep 5, 2026, 04:45 AM |
52 | Mistral Medium 3.5mistral-medium-3-5 | 125.2 tok/s#29 | 0.67 s#17 | 16.64 s#44 | Sep 5, 2026, 04:45 AM |
53 | GPT-5.3 Codexgpt-5-3-codex | 126.2 tok/s#28 | 41.78 s#77 | 41.78 s#67 | Sep 5, 2026, 04:45 AM |
54 | Nex n2 Pronex-n2-pro | 126.3 tok/s#27 | 0.91 s#23 | 16.75 s#45 | Sep 5, 2026, 04:45 AM |
55 | Ring 2 6 1tring-2-6-1t | 126.6 tok/s#26 | 2.06 s#55 | 17.85 s#46 | Sep 5, 2026, 04:45 AM |
56 | DeepSeek V4 Flashdeepseek-v4-flash | 129.7 tok/s#25 | 0.78 s#20 | 16.19 s#43 | Sep 5, 2026, 04:45 AM |
57 | Solar Pro 3solar-pro-3 | 140.9 tok/s#24 | 1.14 s#31 | 15.34 s#42 | Sep 5, 2026, 04:45 AM |
58 | Qwen3.6 35B A3Bqwen3-6-35b-a3b | 141.5 tok/s#23 | 1.09 s#29 | 39.22 s#65 | Sep 5, 2026, 04:45 AM |
59 | Nemotron 3 Supernemotron-3-super-120b-a12b | 141.9 tok/s#22 | 1.13 s#30 | 15.23 s#41 | Sep 5, 2026, 04:45 AM |
60 | Granite 4.2 8Bgranite-4-2-8b | 143.6 tok/s#21 | 0.22 s#1 | 14.15 s#39 | Sep 5, 2026, 04:45 AM |
61 | Qwen3.5-122B-A10Bqwen3-5-122b-a10b | 151.8 tok/s#20 | 1.00 s#27 | 1.00 s#16 | Sep 5, 2026, 04:45 AM |
62 | Qwen3.5-35B-A3Bqwen3-5-35b-a3b | 163.0 tok/s#19 | 1.19 s#36 | 1.19 s#19 | Sep 5, 2026, 04:45 AM |
63 | Nemotron 3 Ultranemotron-3-ultra-550b-a55b | 163.4 tok/s#18 | 1.23 s#39 | 15.16 s#40 | Sep 5, 2026, 04:45 AM |
64 | gpt-oss-120bgpt-oss-120b | 171.7 tok/s#17 | 0.51 s#10 | 12.15 s#37 | Sep 5, 2026, 04:45 AM |
65 | Muse Spark 1.3muse-spark-1-3 | 176.5 tok/s#16 | 18.40 s#72 | 29.73 s#58 | Sep 5, 2026, 04:45 AM |
66 | Qwen3 Next 80B A3B Instructqwen3-next-80b-a3b-instruct | 184.0 tok/s#15 | 1.06 s#28 | 1.06 s#17 | Sep 5, 2026, 04:45 AM |
67 | gpt-oss-20bgpt-oss-20b | 200.2 tok/s#14 | 0.57 s#13 | 10.55 s#34 | Sep 5, 2026, 04:45 AM |
68 | Ling 3.0 Tinyling-3-0-tiny | 206.2 tok/s#13 | 1.90 s#54 | 11.60 s#35 | Sep 5, 2026, 04:45 AM |
69 | Gemini 3.6 Flashgemini-3-6-flash | 206.9 tok/s#12 | 13.81 s#71 | 13.81 s#38 | Sep 5, 2026, 04:45 AM |
70 | Muse Spark 1.2muse-spark-1-2 | 225.8 tok/s#11 | 11.21 s#69 | 20.07 s#48 | Sep 5, 2026, 04:45 AM |
71 | Qwen3.5 Omni Flashqwen3-5-omni-flash | 230.8 tok/s#10 | 0.96 s#26 | 0.96 s#15 | Sep 5, 2026, 04:45 AM |
72 | Gemini 3.5 Flashgemini-3-5-flash | 238.1 tok/s#9 | 0.73 s#19 | 0.73 s#11 | Sep 5, 2026, 04:45 AM |
73 | Nemotron 3.5 Lightningnemotron-3-5-lightning | 294.4 tok/s#8 | 0.52 s#11 | 7.31 s#31 | Sep 5, 2026, 04:45 AM |
74 | Gemini 3.7 Flashgemini-3-7-flash | 294.8 tok/s#7 | 7.14 s#67 | 7.14 s#29 | Sep 5, 2026, 04:45 AM |
75 | Nemotron 3 Nano Omninemotron-3-nano-omni-30b-a3b-reasoning | 327.8 tok/s#6 | 0.56 s#12 | 6.67 s#27 | Sep 5, 2026, 04:45 AM |
76 | Trinity Large Thinkingtrinity-large-thinking | 327.8 tok/s#5 | 1.17 s#33 | 7.27 s#30 | Sep 5, 2026, 04:45 AM |
77 | Ling-3.0-flashling-3-0-flash | 336.0 tok/s#4 | 1.71 s#49 | 7.66 s#32 | Sep 5, 2026, 04:45 AM |
78 | Gemini 3.5 Flash-Litegemini-3-5-flash-lite | 352.1 tok/s#3 | 6.53 s#65 | 6.53 s#26 | Sep 5, 2026, 04:45 AM |
79 | Gemini 3.8 Flashgemini-3-8-flash | 429.2 tok/s#2 | 19.74 s#73 | 19.74 s#47 | Sep 5, 2026, 04:45 AM |
80 | Mercury 2mercury-2 | 907.0 tok/s#1 | 4.15 s#63 | 4.15 s#24 | Sep 5, 2026, 04:45 AM |
