Week of 17 Oct 2026 · 10 labs tracked

Who's winning
the AI race?

One score per lab, built from model rankings, money, and prediction-market odds. Every number links to its source. How it's scored →

👑 NEW #1 THIS WEEK

Anthropic

Takes the top spot from Google DeepMind: 94.4 vs 93.3. Polymarket's 2026 crown odds are +6pp on the week, to 47.5%.

Standings Score 0–100 · arrows are rank change vs 10 Oct 2026 (backfilled)

Pos		Lab · top model	Score	Arena	AA Index	Revenue run-rate	Valuation	2026 #1 odds
1	▲1	AnthropicNEW #1Claude Opus 5.5 (High)	94.4+0.5 · 5/5 data	#31508	57.6#1	$65B	$965B	47.5%+6pp
2	▼1	Google DeepMindGemini 4 Argon (High)	93.3-4.8 · 3/5 data	#11524	52.6#9	—	—	39.5%-3pp
3	—	OpenAIGPT-5.6 Sol (xHigh)	67.4-1.4 · 5/5 data	#211485	52.7#8	$50BCLAIM $70B	$852Btalks $1.4T	7.8%-2.1pp
4	—	Meta Superintelligence LabsMuse Wonder (Max)	64.9+0.2 · 3/5 data	#21509	55.8#4	—	—	2.8%+0.6pp
5	—	Alibaba (Qwen)Qwen3.8-Max	50.7±0 · 4/5 data	#241482	45.4#29	$7.3Bq × 4	—	0.5%±0pp
6	—	xAI (SpaceXAI)Grok 4.20 Beta 1	49.4-0.3 · 5/5 data	#371475	46.4#25	$3.3Bq × 4	$250B	1.7%-0.4pp
7	—	Moonshot AI (Kimi)Kimi K3 (Max)	46.3-0.2 · 5/5 data	#181487	43.6#33	$1BCLAIM $2B	$50B	0.4%-0.2pp
8	—	Z.ai (Zhipu)GLM-5.3 (Max)	44.5+0.1 · 3/5 data	#281479	44.8#31	—	—	0.5%+0.1pp
9	—	DeepSeekDeepSeek-V4.1-Flash (Max)	41.2±0 · 5/5 data	#381475	39.5#47	$1BEST	$51.8B	0.5%+0.1pp
10	—	Mistral AIMistral Large 4	24.6±0 · 5/5 data	#1111432	38.4#49	$1B	$24B	0.3%±0pp

"—" means no reliable public number, so it is left out of the score rather than guessed. Big-tech labs (Google, Meta) don't report lab-level revenue or valuation. CLAIM = a higher figure the company or its investors put out. EST = third-party estimate.