SmophyAI

Published daily · Next update 02:00 UTC

Themed ranking

Fastest and most reliable

Ranked by real measured throughput (tokens/sec) at each model's best-performing provider, tracked daily.

As of November 7, 2026, Qwen: Qwen3.8 27B (batch) ranks #1 for fastest and most reliable at throughput (tok/s) 835 tok/s.

  1. #1 Qwen: Qwen3.8 27B (batch) - Throughput (tok/s) 835 tok/s
  2. #2 OpenAI: gpt-oss-120b (batch) - Throughput (tok/s) 813 tok/s
  3. #3 Z.ai: GLM 5.2 - Throughput (tok/s) 528 tok/s
  4. #4 OpenAI: gpt-oss-20b (batch) - Throughput (tok/s) 453 tok/s
  5. #5 NVIDIA: Nemotron 3.5 Super VL (free) - Throughput (tok/s) 310 tok/s
  6. #6 Google: Gemma 4 31B (free) - Throughput (tok/s) 304 tok/s
  7. #7 NVIDIA: Nemotron 3.5 Lightning (free) - Throughput (tok/s) 269 tok/s
  8. #8 MiniMax: MiniMax M2.7 - Throughput (tok/s) 228 tok/s
  9. #9 inclusionAI: Ling 3.0 Flash Fin - Throughput (tok/s) 213 tok/s
  10. #10 DeepSeek: DeepSeek V4 Flash 0423 - Throughput (tok/s) 213 tok/s
50 eligible models, sorted by throughput (tok/s)
#ModelThroughput (tok/s)
01Qwen: Qwen3.8 27B (batch)835 tok/s
02OpenAI: gpt-oss-120b (batch)813 tok/s
03Z.ai: GLM 5.2528 tok/s
04OpenAI: gpt-oss-20b (batch)453 tok/s
05NVIDIA: Nemotron 3.5 Super VL (free)310 tok/s
06Google: Gemma 4 31B (free)304 tok/s
07NVIDIA: Nemotron 3.5 Lightning (free)269 tok/s
08MiniMax: MiniMax M2.7228 tok/s
09inclusionAI: Ling 3.0 Flash Fin213 tok/s
10DeepSeek: DeepSeek V4 Flash 0423213 tok/s
11Z.ai: GLM 5.3 (batch)201 tok/s
12MiniMax: MiniMax M3.1 Flash199 tok/s
13MiniMax: MiniMax M3 (batch)196 tok/s
14Qwen: Qwen3.6 35B A3B196 tok/s
15DeepSeek: DeepSeek V4.1 Flash (batch)193 tok/s
16Anthropic: Claude Haiku 5.5 (batch)189 tok/s
17Google: Gemini 2.5 Flash Lite181 tok/s
18Google: Gemini 3.1 Flash Lite (batch)178 tok/s
19StepFun: Step 5 Preview168 tok/s
20Thinking Machines: Inkling164 tok/s
21Z.ai: GLM 5.3 Flash (batch)151 tok/s
22Google: Gemini 2.5 Flash145 tok/s
23OpenAI: GPT-6 Luna (batch)143 tok/s
24Google: Gemma 4 26B A4B (free)143 tok/s
25Google: Gemini 3.5 Flash (batch)131 tok/s
26OpenAI: GPT-5.6 Sol (batch)131 tok/s
27Mistral: Mistral Nemo130 tok/s
28Meta: Muse Spark 1.2 Contributor126 tok/s
29DeepSeek: DeepSeek V4 Pro 0423122 tok/s
30MoonshotAI: Kimi K3 (batch)121 tok/s
31OpenAI: GPT-5.6 Luna (batch)115 tok/s
32NVIDIA: Nemotron 3 Super115 tok/s
33NVIDIA: Nemotron 3 Ultra (free)114 tok/s
34Poolside: Laguna XS 2.1 (free)114 tok/s
35Qwen: Qwen3.8 Flash112 tok/s
36Anthropic: Claude Sonnet 5.5 (batch)111 tok/s
37OpenAI: GPT-5.6 Terra (batch)109 tok/s
38Xiaomi: MiMo-V2.6-Pro109 tok/s
39DeepSeek: DeepSeek V3.2108 tok/s
40Meta: Muse Spark 1.2105 tok/s
41Tencent: Hy4 preview103 tok/s
42inclusionAI: Ling-2.6-flash101 tok/s
43OpenAI: GPT-5.6 Luna Pro (batch)99 tok/s
44OpenAI: GPT-5 Mini (batch)99 tok/s
45Poolside: Laguna S 2.1 (free)97 tok/s
46StepFun: Step 3.7 Flash96 tok/s
47Nex AGI: Nex-N2-Mini94 tok/s
48Meta: Muse Spark 1.3 Contributor92 tok/s
49Anthropic: Claude Opus 5.5 (batch)89 tok/s
50DeepSeek: DeepSeek V4 Flash Vision Exp88 tok/s

Source: OpenRouter (openrouter.ai/rankings), as of November 7, 2026.

Benchmark scores: Artificial Analysis (artificialanalysis.ai) via OpenRouter (openrouter.ai/rankings).

Methodology: www.smophy.ai/benchmark/methodology

Token counts originate from each provider's own tokenizer and are not directly comparable across providers.