OpenAI vs Anthropic vs Google DeepMind
The three highest-profile labs in the race, side by side: score breakdown, model ranks, money and Polymarket odds, every figure sourced, and who leads each category.
Data as of the week of 26 Oct 2026 (updated 26 Oct 2026, 08:09, Bucharest time). Every figure links to its source; see the methodology.
Score, models, money and odds
| Metric | OpenAI | Anthropic | Google DeepMind | Leads |
|---|---|---|---|---|
| Composite score | 68.6/100 | 97.2/100 | 86.9/100 | Anthropic |
| Arena text rank | #21 GPT-5.6 Sol (xHigh) | #2 Claude Opus 5.5 (High) | #1 Gemini 4 Argon (High) | Google DeepMind |
| AA Intelligence Index | 52.7 GPT-6 Astra (Max) | 57.6 Claude Opus 5.5 (Max, Default Fallback) | 52.6 Gemini 4 Argon (High) | Anthropic |
| Revenue run-rate | $50B annualized run-rate, end of September (net of partner share) | $67.2B Q3 2026 revenue × 4 (computed; $16.8B quarter told to investors, gross accounting) | n/a | Anthropic |
| Valuation (latest closed) | $852B post-money, $122B round | $965B post-money, Series H | n/a | Anthropic |
| Total funding (not scored) | $178B | $141B | n/a | OpenAI |
| Polymarket 2026 #1 odds | 5.7% −1.6pp vs 19 Oct 2026 | 53.5% +4pp vs 19 Oct 2026 | 32.5% −7pp vs 19 Oct 2026 | Anthropic |
| Weekly changes vs 19 Oct 2026 | ||||
| Score change | +1.7 | +2.8 | −5.4 | Anthropic |
| Rank change | — | — | — | tie: OpenAI, Anthropic, Google DeepMind |
| Odds change (pp) | −1.6pp | +4pp | −7pp | Anthropic |
Revenue run-rates use each company's own accounting and aren't perfectly comparable: Anthropic books some cloud-partner revenue gross and OpenAI books it net (EST = third-party estimate). Funding and compute are shown but not scored. See the methodology.
Compute (not scored)
OpenAI
- Planned capacity by 2030 (told investors): 30 GW (9 Sept 2026 · AI-RTZ (opens in new tab))
- Compute spend forecast through 2030: ~$750B (9 Sept 2026 · AI-RTZ (The Information) (opens in new tab))
Anthropic
- Compute contracted since Oct 2025: ≥14.8 GW (6 Sept 2026 · The Information via AI-RTZ (opens in new tab))
- Anthropic-Amazon New Carlisle, current IT power (est.): ~910 MW (15 Oct 2026 · Epoch AI (opens in new tab))
- Fixed/minimum compute & infrastructure commitments: ~$265.6B (30 Jun 2026 · Anthropic S-1, Note 13 (SEC EDGAR) (opens in new tab))
Google DeepMind
- Alphabet 2026 capex guidance (company-wide): $195–205B (22 Jul 2026 · CNBC (Q2 2026 earnings) (opens in new tab))
Score history
| Snapshot | OpenAI | Anthropic | Google DeepMind |
|---|---|---|---|
| 10 Oct 2026 backfilled | #3 · 68.8 | #2 · 93.9 | #1 · 98.1 |
| 17 Oct 2026 | #3 · 67.3 | #1 · 94.4 | #2 · 93.3 |
| 19 Oct 2026 | #3 · 66.9 | #1 · 94.4 | #2 · 92.3 |
| 22 Oct 2026 | #3 · 66.1 | #2 · 94.4 | #1 · 96 |
| 26 Oct 2026 | #3 · 68.6 | #1 · 97.2 | #2 · 86.9 |
Backfilled snapshots use real historical Polymarket odds; Arena and AA were held at current values and financials were limited to facts dated on or before that day.
FAQ
Which leads the AI race right now: OpenAI, Anthropic or Google DeepMind?
Anthropic on the composite score (97.2/100), week of 26 Oct 2026.
Which of the three has the highest revenue run-rate?
Anthropic, $67.2B. Anthropic books some cloud-partner revenue gross and OpenAI books it net, so the two aren't perfectly comparable.
Which do Polymarket traders favour for the best AI model at the end of 2026?
Anthropic, at 53.5%.
Which has the top-ranked model on the Arena leaderboard?
Google DeepMind, #1 as of 22 Oct 2026.
Why does Google DeepMind have no revenue or valuation figure?
This is a division of a public company. Alphabet does not disclose lab-level revenue or valuation, so those score components are excluded and the remaining weights are rescaled. Its score is computed on the remaining components (3/5); see the methodology.