Vikshy is not being updated. Updates were paused on 20 October 2026, so the news, the Score, prices and rankings here are out of date and will stay as they are. Check the labs' own sites for anything newer.
The Vikshy Score · closed 06:00 UTC, 28 August 2026 · method v2.3 · a frozen close

The Score on 28 August 2026.

On 28 August 2026, OpenAI leads the Vikshy Score at 75.6, ahead of Google DeepMind (73.3) and Anthropic (70.4). Google DeepMind is rising fastest (▲ 4.6, after Google publishes TimesFM 3.0 PyTorch on Hugging Face); Anthropic is falling fastest (▼ 2.1, after Tests find Anthropic's Opus 4.6 can bypass explicit content limits).

LeaderOpenAI75.62.4 ahead of Google Rising fastest, 30 daysGoogle DeepMind▲ 4.6After Google publishes TimesFM 3.0 PyTorch on Hugging Face Falling fastest, 30 daysAnthropic▼ 2.1After Tests find Anthropic's Opus 4.6 can bypass explicit content limits
#LabScoreStrengthDirection30 daysConfidence
1OpenAIAfter OpenAI announces Astra, a next major model family75.675.5▲ 0.82Google DeepMindAfter Google publishes TimesFM 3.0 PyTorch on Hugging Face73.372.4▲ 4.63AnthropicAfter Tests find Anthropic's Opus 4.6 can bypass explicit content limits70.470.8▼ 2.14AlibabaAfter Alibaba announces Qwen3.8-Max flagship model with 2.4 trillion parameters63.462.6▲ 4.35Z.aiAfter Z.ai releases GLM-5.3 for frontier coding63.362.6▲ 4.06MetaAfter Meta introduces Muse Spark 1.258.157.6▲ 2.57xAIAfter xAI releases Grok 4.6, matching GPT-5.6 Sol on intelligence index56.556.0▲ 3.08DeepSeekAfter DeepSeek ships improved V4 Pro and open-sources its agent software55.354.9▲ 1.99MistralAfter Mistral launches Shieldstral open-weight moderation model53.653.4▲ 0.910TencentAfter Tencent publishes Hy4-preview model on Hugging Face52.852.0▲ 4.311MoonshotNo material launch or update in 30 days51.151.3▼ 1.012MiniMaxAfter MiniMax releases MiniMax-Music3 music generation model on Hugging Face50.550.3▲ 1.313CohereAfter Cohere releases North Small Translate 1.0 fp8 on Hugging Face41.541.1▲ 2.2

Score is Strength today plus 0.18 × Direction over 30 days. Confidence is how much of a lab's evidence is measured and fresh. How the Score works

Reconstructed Rebuilt with the live code from evidence dated on or before 28 August 2026. How

Strength and Direction

Across: where each lab stands today. Up: which way it has moved over the last 30 days. Filled dots are rising.

How the Score works

The Vikshy Score is a daily, open ranking of the leading AI labs, built only from public evidence: how strong each lab is today, which way it is heading, and exactly why. Strength is where each lab stands, from eight measured pillars; Direction is where it is heading, from the last 30 days of developments. 13 labs ranked, 4 on the watchlist. The Score closes daily at 06:00 UTC; big launches and outages move Direction live in between.

History

Every close is kept with the method version it used. Past numbers are never rewritten; corrections are logged in the open.

Lab13 Oct14 Oct15 Oct16 Oct17 Oct18 Oct19 Oct
OpenAI78.878.878.878.878.978.978.8
Google75.275.575.975.875.775.775.7
Anthropic79.579.579.679.979.979.879.8
Alibaba61.861.661.761.661.661.361.6
Z.ai58.158.158.358.258.258.258.2
Meta60.960.961.161.962.162.062.0
xAI58.658.658.658.758.758.858.7
DeepSeek58.158.159.057.657.558.558.5
Mistral56.356.356.555.955.855.855.8
Tencent51.551.551.650.150.650.650.6
Moonshot50.450.450.550.550.550.551.0
MiniMax45.945.946.345.945.745.745.7
Cohere41.741.841.941.641.641.641.6

Share the Score

Every way the Score leaves this site carries the date, the method version and a link home. Shared numbers are always a daily close, never a live reading.