SmophyAI

Published daily · Next update 02:00 UTC

Daily snapshot · immutable permalink

AI model usage rankings - September 20, 2026

As published on September 20, 2026, Z.ai: GLM 5.3 Flash (batch) led real OpenRouter usage share at 14.5%. Xiaomi: MiMo-V2.5-Pro, Qwen: Qwen3.8 27B (batch), SpaceXAI: Grok 4.6 (batch) entered the leaderboard for the first time. Qwen: Qwen3.8 Max moved 16 places down since September 19, 2026.

  1. #1 Z.ai: GLM 5.3 Flash (batch) - 14.5% usage share
  2. #2 DeepSeek: DeepSeek V4.1 Flash (batch) - 14.4% usage share
  3. #3 Tencent: Hy4 preview - 11.4% usage share
  4. #4 DeepSeek: DeepSeek V4 Flash 0423 - 8.9% usage share
  5. #5 OpenAI: GPT-5.6 Luna (batch) - 7.0% usage share
  6. #6 Xiaomi: MiMo-V2.5 - 5.1% usage share
  7. #7 NVIDIA: Nemotron 3 Ultra (free) - 4.5% usage share
  8. #8 Tencent: Hy3 - 2.8% usage share
  9. #9 Z.ai: GLM 5.3 (batch) - 1.8% usage share
  10. #10 Meta: Muse Spark 1.3 Contributor - 1.7% usage share
Change since September 19, 2026
#ModelUsage shareRank change
01Z.ai: GLM 5.3 Flash (batch)14.5%+1 (-0.6pts)
02DeepSeek: DeepSeek V4.1 Flash (batch)14.4%-1 (-2.9pts)
03Tencent: Hy4 preview11.4%+1 (+3.5pts)
04DeepSeek: DeepSeek V4 Flash 04238.9%-1 (-2.3pts)
05OpenAI: GPT-5.6 Luna (batch)7.0%no change
06Xiaomi: MiMo-V2.55.1%no change
07NVIDIA: Nemotron 3 Ultra (free)4.5%no change
08Tencent: Hy32.8%no change
09Z.ai: GLM 5.3 (batch)1.8%no change
10Meta: Muse Spark 1.3 Contributor1.7%no change
11Upstage: Solar Pro 41.5%+1 (+0.1pts)
12Google: Gemini 3.8 Flash (batch)1.5%+1 (+0.1pts)
13Z.ai: GLM 5.21.4%+1 (+0.3pts)
14DeepSeek: DeepSeek V4 Pro 04231.3%-3 (-0.2pts)
15OpenAI: GPT-5.6 Sol (batch)1.1%+4 (+0.2pts)
16MiniMax: MiniMax M3 (batch)1.0%no change
17MoonshotAI: Kimi K3 (batch)1.0%-2 (-0.1pts)
18Poolside: Laguna S 2.1 (free)1.0%-1 (0pts)
19inclusionAI: Ling 3.0 Flash Fin1.0%-1 (+0.1pts)
20Anthropic: Claude Sonnet 5 (batch)0.8%+1 (0pts)
21Anthropic: Claude Opus 5 (batch)0.7%+1 (+0.1pts)
22OpenAI: GPT-6 Astra (batch)0.6%-2 (-0.2pts)
23Nex AGI: Nex-N2.5-Pro0.6%+3 (+0.2pts)
24Google: Gemini 3 Flash Preview (batch)0.5%no change
25Google: Gemini 2.5 Flash Lite0.5%-2 (-0.1pts)
26inclusionAI: Ling 3.0 Flash VL0.5%-1 (-0.1pts)
27NVIDIA: Nemotron 3.5 Lightning (free)0.4%no change
28OpenAI: GPT-5.6 Terra (batch)0.4%+6 (+0.1pts)
29Google: Gemini 3.7 Flash (batch)0.4%-1 (0pts)
30Qwen: Qwen3.8 Flash0.4%+3 (0pts)
31Qwen: Qwen3.7 Flash0.4%no change
32OpenAI: gpt-oss-120b (batch)0.3%+3 (0pts)
33Thinking Machines: Inkling0.3%+7 (+0.1pts)
34Anthropic: Claude Fable 5.1 (batch)0.3%-4 (0pts)
35DeepSeek: DeepSeek V3.20.3%+1 (0pts)
36Google: Gemini 3.1 Flash Lite (batch)0.3%+1 (0pts)
37Mistral: Mistral Nemo0.3%+2 (0pts)
38Google: Gemini 2.5 Flash0.3%-6 (0pts)
39Google: Gemma 4 31B (free)0.3%+3 (0pts)
40NVIDIA: Nemotron 3 Super0.3%+1 (0pts)
41Xiaomi: MiMo-V2.5-Pro New0.3%-
42Qwen: Qwen3.8 27B (batch) New0.2%-
43inclusionAI: Ling 3.0 Flash Sante0.2%+2 (0pts)
44Google: Gemma 4 26B A4B (free)0.2%-1 (0pts)
45Qwen: Qwen3.8 Max0.2%-16 (-0.1pts)
46SpaceXAI: Grok 4.6 (batch) New0.2%-

This page is a permanent, unchanging record of real OpenRouter usage share as it stood on September 20, 2026. Later corrections to the data pipeline appear as new dated snapshots - this page's figures never change.

Source: OpenRouter (openrouter.ai/rankings), as of September 20, 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.