SmophyAI

Published daily · Next update 02:00 UTC

Daily snapshot · immutable permalink

AI model usage rankings - September 3, 2026

As published on September 3, 2026, Tencent: Hy4 preview led real OpenRouter usage share at 16.1%. Meta: Muse Spark 1.3 Contributor, Meta: Muse Spark 1.3, MoonshotAI: Kimi K2.6 entered the leaderboard for the first time. Google: Gemini 3.8 Flash (batch) moved 23 places up since September 2, 2026.

  1. #1 Tencent: Hy4 preview - 16.1% usage share
  2. #2 DeepSeek: DeepSeek V4 Flash 0423 - 14.5% usage share
  3. #3 Z.ai: GLM 5.3 Flash (batch) - 10.2% usage share
  4. #4 OpenAI: GPT-5.6 Luna (batch) - 7.8% usage share
  5. #5 MiniMax: MiniMax M3 - 6.1% usage share
  6. #6 Tencent: Hy3 - 4.1% usage share
  7. #7 Z.ai: GLM 5.3 (batch) - 3.1% usage share
  8. #8 NVIDIA: Nemotron 3 Ultra (free) - 2.8% usage share
  9. #9 Google: Gemini 3.7 Flash (batch) - 2.2% usage share
  10. #10 Z.ai: GLM 5.2 (free) - 2.1% usage share
Change since September 2, 2026
#ModelUsage shareRank change
01Tencent: Hy4 preview16.1%+2 (+3.8pts)
02DeepSeek: DeepSeek V4 Flash 042314.5%no change
03Z.ai: GLM 5.3 Flash (batch)10.2%+1 (+0.5pts)
04OpenAI: GPT-5.6 Luna (batch)7.8%-3 (-7.6pts)
05MiniMax: MiniMax M36.1%no change
06Tencent: Hy34.1%no change
07Z.ai: GLM 5.3 (batch)3.1%+2 (+0.5pts)
08NVIDIA: Nemotron 3 Ultra (free)2.8%-1 (-0.2pts)
09Google: Gemini 3.7 Flash (batch)2.2%-1 (-0.5pts)
10Z.ai: GLM 5.2 (free)2.1%+3 (+0.4pts)
11OpenAI: GPT-5.6 Sol (batch)2.1%+1 (+0.2pts)
12DeepSeek: DeepSeek V4 Pro 04231.9%-1 (-0.1pts)
13MoonshotAI: Kimi K3 (batch)1.8%+1 (+0.2pts)
14Xiaomi: MiMo-V2.51.7%-4 (-0.8pts)
15Anthropic: Claude Opus 5 (batch)1.5%no change
16Google: Gemini 3.8 Flash (batch)1.5%+23 (+1.2pts)
17Anthropic: Claude Sonnet 5 (batch)1.1%-1 (0pts)
18Anthropic: Claude Sonnet 4.6 (batch)1.1%+3 (+0.4pts)
19Poolside: Laguna S 2.1 (free)1.1%-2 (0pts)
20NVIDIA: Nemotron 3.5 Lightning (free)0.8%-1 (+0.1pts)
21Upstage: Solar Pro 40.8%-3 (+0.1pts)
22Meta: Muse Spark 1.3 Contributor New0.8%-
23Google: Gemini 3 Flash Preview (batch)0.7%-3 (0pts)
24inclusionAI: Ling 3.0 Flash Fin (free)0.6%-2 (0pts)
25Google: Gemini 2.5 Flash Lite (batch)0.6%-2 (0pts)
26OpenAI: GPT-5.6 Terra (batch)0.5%-2 (0pts)
27MiniMax: MiniMax M2.70.5%-2 (0pts)
28Google: Gemini 2.5 Flash (batch)0.4%no change
29Anthropic: Claude Fable 5.1 (batch)0.4%+2 (0pts)
30StepFun: Step 3.7 Flash0.4%-4 (-0.1pts)
31Google: Gemini 3.1 Flash Lite (batch)0.4%-1 (0pts)
32DeepSeek: DeepSeek V4 Flash Vision Exp0.4%-3 (-0.1pts)
33DeepSeek: DeepSeek V3.20.4%+1 (0pts)
34OpenAI: gpt-oss-120b0.3%-7 (-0.1pts)
35OpenAI: gpt-oss-20b (batch)0.3%-3 (0pts)
36OpenAI: GPT-5.6 Luna Pro (batch)0.3%+2 (0pts)
37Google: Gemma 4 26B A4B (free)0.3%+3 (0pts)
38NVIDIA: Nemotron 3 Super (free)0.3%+7 (+0.1pts)
39Google: Gemma 4 31B (free)0.3%-3 (0pts)
40Anthropic: Claude Opus 4.8 (batch)0.3%-5 (0pts)
41SpaceXAI: Grok 4.60.3%-4 (0pts)
42Meta: Muse Spark 1.3 New0.2%-
43Xiaomi: MiMo-V2.5-Pro0.2%-2 (0pts)
44Qwen: Qwen3.8 Max0.2%-1 (0pts)
45Anthropic: Claude Haiku 4.5 (batch)0.2%-3 (0pts)
46MoonshotAI: Kimi K2.6 New0.2%-
47Qwen: Qwen3.8 27B (free)0.2%-3 (0pts)

This page is a permanent, unchanging record of real OpenRouter usage share as it stood on September 3, 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 3, 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.