whatmodel.sh / how far behind is your GPU

How to use it

  1. Find your line by your graphics card's memory: 8 GB, 24 GB, or any size if you rent a server.
  2. Point at any month to see the best model on each line then, and how far behind closed models it was.
  3. Click a model's name for its scores, every source, and the command to run it.
  4. For your exact machine, enter its GPU memory and RAM on the leaderboard: 24 GB card · 8 GB card. Each model shows the version that fits and the command to paste.

How it's worked out

Each new best closed model, and how long until you could run its equal

Closed modelreleasedscoreopen weights, any sizefits a 24 GB GPUfits in 8 GB
GPT-4 (0314)Mar 202353.615 months · DeepSeek-Coder-V2 Instruct (0614)18 months · Qwen2.5 32B Instruct24 months · Qwen2.5-Omni 7B
GPT-4 Turbo Preview (1106)Nov 202354.68 months · Llama 3.1 405B Instruct10 months · Qwen2.5 32B Instruct16 months · Qwen2.5-Omni 7B
GPT-4 Turbo (2024-04-09)Apr 202456.17 months · Qwen2.5 Coder 32B Instruct7 months · Qwen2.5 Coder 32B Instruct11 months · Qwen2.5-Omni 7B
GPT-4o (2024-05-13)May 202457.67 months · DeepSeek V3 (Dec '24)10 months · QwQ-32B22 months · Qwen3.5 9B
GPT-4o (2024-08-06)Aug 202457.74 months · DeepSeek V3 (Dec '24)7 months · QwQ-32B19 months · Qwen3.5 9B
o1-previewSep 202462.98 months · DeepSeek R1 052817 months · Qwen3.5 27Bnot yet (26 months so far)
o1 (2024-12-17)Dec 202465.37 months · Qwen3 235B-A22B Thinking 250714 months · Qwen3.5 27Bnot yet (23 months so far)
Gemini 2.5 Pro Experimental (03-25)Mar 2025679 months · GLM-4.711 months · Qwen3.5 27Bnot yet (20 months so far)
o3Apr 202568.19 months · Kimi K2.516 months · Qwen3.8 27Bnot yet (19 months so far)
Gemini 2.5 Pro Preview (06-05)Jun 202569.110 months · Kimi K2.614 months · Qwen3.8 27Bnot yet (17 months so far)
GPT-5Aug 202570.88 months · Kimi K2.612 months · Qwen3.8 27Bnot yet (15 months so far)
GPT-5 ProOct 202571.79 months · Kimi K3not yet (13 months so far)not yet (13 months so far)
GPT-5.2 ProDec 202575.78 months · Qwen3.8 Max (2.4T-A95B)not yet (11 months so far)not yet (11 months so far)
GPT-5.4 ProMar 202678.2not yet (8 months so far)not yet (8 months so far)not yet (8 months so far)
Claude Fable 5Jun 202678.6not yet (5 months so far)not yet (5 months so far)not yet (5 months so far)
Claude Opus 5Jul 202679.8not yet (4 months so far)not yet (4 months so far)not yet (4 months so far)
GPT-6 AstraSep 202684.4not yet (2 months so far)not yet (2 months so far)not yet (2 months so far)
Claude Fable 5.2Nov 202688.5not yet (0 months so far)not yet (0 months so far)not yet (0 months so far)

Every result links to its source on the model's page. The data (CC BY-SA 4.0) and code (MIT) are on GitHub. See also the open vs closed race.