SmophyAI

Published daily · Next update 02:00 UTC

Daily snapshot · immutable permalink

AI model usage rankings - August 22, 2026

As published on August 22, 2026, Ox Alpha led real OpenRouter usage share at 32.3%. Qwen: Qwen3.7 Flash, OpenAI: gpt-oss-20b entered the leaderboard for the first time. Google: Gemini 3.6 Flash (batch) moved 24 places down since August 21, 2026.

  1. #1 Ox Alpha - 32.3% usage share
  2. #2 DeepSeek: DeepSeek V4 Flash 0423 - 14.7% usage share
  3. #3 Xiaomi: MiMo-V2.5 - 10.0% usage share
  4. #4 Tencent: Hy3 - 5.4% usage share
  5. #5 NVIDIA: Nemotron 3 Ultra (free) - 5.2% usage share
  6. #6 OpenAI: GPT-5.6 Luna (batch) - 3.0% usage share
  7. #7 DeepSeek: DeepSeek V4 Pro 0423 - 2.6% usage share
  8. #8 Z.ai: GLM 5.2 (free) - 2.3% usage share
  9. #9 Google: Gemini 3.7 Flash (batch) - 1.6% usage share
  10. #10 Poolside: Laguna S 2.1 (free) - 1.4% usage share
Change since August 21, 2026
#ModelUsage shareRank change
01Ox Alpha32.3%+1 (+17.9pts)
02DeepSeek: DeepSeek V4 Flash 042314.7%-1 (-4.2pts)
03Xiaomi: MiMo-V2.510.0%no change
04Tencent: Hy35.4%no change
05NVIDIA: Nemotron 3 Ultra (free)5.2%no change
06OpenAI: GPT-5.6 Luna (batch)3.0%no change
07DeepSeek: DeepSeek V4 Pro 04232.6%+1 (-0.1pts)
08Z.ai: GLM 5.2 (free)2.3%-1 (-0.7pts)
09Google: Gemini 3.7 Flash (batch)1.6%no change
10Poolside: Laguna S 2.1 (free)1.4%no change
11MiniMax: MiniMax M3 (free)1.3%+2 (-0.2pts)
12MoonshotAI: Kimi K31.2%-1 (-0.3pts)
13Claude Opus 5 (batch)1.0%+1 (-0.3pts)
14NVIDIA: Nemotron 3.5 Lightning (free)1.0%+2 (-0.1pts)
15OpenAI: GPT-5.6 Sol (batch)1.0%-3 (-0.4pts)
16Anthropic: Claude Sonnet 5 (batch)0.8%-1 (-0.4pts)
17Google: Gemini 3 Flash Preview (batch)0.7%no change
18Z.ai: GLM 5.30.7%no change
19Google: Gemini 2.5 Flash Lite (batch)0.6%no change
20Upstage: Solar Pro 40.6%+1 (-0.1pts)
21OpenAI: gpt-oss-120b0.5%+7 (0pts)
22Anthropic: Claude Sonnet 4.6 (batch)0.5%+1 (-0.1pts)
23SpaceXAI: Grok 4.60.5%+3 (-0.1pts)
24DeepSeek: DeepSeek V3.20.4%+8 (0pts)
25StepFun: Step 3.7 Flash0.4%+5 (-0.1pts)
26Xiaomi: MiMo-V2.5-Pro0.4%-1 (-0.1pts)
27OpenAI: GPT-5.6 Terra (batch)0.4%-5 (-0.3pts)
28NVIDIA: Nemotron 3 Super (free)0.4%+8 (0pts)
29Google: Gemini 2.5 Flash (batch)0.4%-2 (-0.2pts)
30Google: Gemini 3.1 Flash Lite (batch)0.3%-1 (-0.2pts)
31Google: Gemma 4 26B A4B (free)0.3%+6 (0pts)
32Google: Gemma 4 31B (free)0.3%+2 (-0.1pts)
33Qwen: Qwen3.8 Max0.3%-9 (-0.3pts)
34Anthropic: Claude Opus 4.8 (batch)0.3%-1 (-0.2pts)
35OpenAI: GPT-5.5 (batch)0.3%+3 (0pts)
36OpenAI: GPT-5.6 Luna Pro (batch)0.2%-5 (-0.2pts)
37Mistral: Mistral Nemo0.2%+8 (0pts)
38MiniMax: MiniMax M2.7 (free)0.2%+5 (0pts)
39OpenAI: GPT-4o-mini (batch)0.2%+2 (0pts)
40Google: Gemini 3.1 Pro Preview (batch)0.2%+4 (0pts)
41Anthropic: Claude Fable 5 (batch)0.2%-2 (-0.1pts)
42Anthropic: Claude Haiku 4.5 (batch)0.2%no change
43OpenAI: GPT-5 Mini (batch)0.2%-3 (-0.1pts)
44Google: Gemini 3.6 Flash (batch)0.2%-24 (-0.6pts)
45MoonshotAI: Kimi K2.60.1%+1 (-0.1pts)
46Qwen: Qwen3.7 Flash New0.1%-
47OpenAI: gpt-oss-20b New0.1%-

This page is a permanent, unchanging record of real OpenRouter usage share as it stood on August 22, 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 August 22, 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.