The model averaged -3.87% in down environments and +0.93% in up environments, a 4.8 percentage-point gap.
Market environments group resolved rounds by the S&P 500 return over the same weekly or monthly window. Models are compared only on shared rounds; high confidence requires at least six observations and stable leadership.
Medium confidenceMath: deterministicData through Jul 9, 2026
Average Return Gap
+4.80%
Down Average Return
-3.87%
Up Average Return
+0.93%
Confidence CalibrationAs of Jul 10
All resolved official results36 resolved rounds181 scored resultsMedian confidence 0.55Resolved history
Across resolved official results, submissions at or above the median confidence of 0.55 averaged -0.57%, while lower-confidence submissions averaged -0.36%.
Confidence is the model's own 0-1 self-reported confidence at submission time, compared with later realized returns.
High confidenceMath: deterministicData through Jul 10, 2026
#5Rank in set-19.7CapitalBench Score-37.8%Total return23Shared rounds
Current Weekly Benchmark
CapitalBench Score
-19.70.050.0100.0
-4.5
-7.0
-8.4
-17.3
-19.7
-0.2
100.0
Claude Opus 4.8
Grok 4.3
Claude Opus 4.7
GPT-5.5
Gemini 3.1 ProThis model
S&PS&P 500
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.hindsight best asset
Higher benchmark score is better.
Claude Opus 4.8Anthropic · 23/23 scored rounds
-4.5
Grok 4.3xAI · 23/23 scored rounds
-7.0
Claude Opus 4.7Anthropic · 23/23 scored rounds
-8.4
GPT-5.5OpenAI · 23/23 scored rounds
-17.3
Gemini 3.1 ProGoogle · 23/23 scored rounds
-19.7
S&PS&P 500S&P 500 · 23/23 scored rounds
-0.2
MAXMax possibleHindsight ceiling, not a model portfolio
What is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio. 100.0
23 shared resolved rounds5 equal-run models rankedQualified at 6+ shared roundsNewest included round: CB-2026-07-02-1W
Return context
Average Return Details
Average portfolio return across the same finished rounds.
Claude Opus 4.8
-0.38%
Grok 4.3
-0.58%
Claude Opus 4.7
-0.70%
GPT-5.5
-1.45%
Gemini 3.1 Pro
-1.64%
S&PS&P 500
-0.02%
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.
8.33%
Fairness rule: every ranked model completed every included round. A missed round is excluded from this set for everyone.
Monthly · Current
May 28, 2026 roster
Monthly comparison set that adds Claude Opus 4.8 to the established model roster.
#4Rank in set-9.9CapitalBench Score-16.5%Total return8Shared rounds
Current Monthly Benchmark
CapitalBench Score
-15.60.050.0100.0
0.5
0.2
-0.4
-9.9
-15.6
-2.6
100.0
Claude Opus 4.8
Grok 4.3
Claude Opus 4.7
Gemini 3.1 ProThis model
GPT-5.5
S&PS&P 500
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.hindsight best asset
Higher benchmark score is better.
Claude Opus 4.8Anthropic · 8/8 scored rounds
0.5
Grok 4.3xAI · 8/8 scored rounds
0.2
Claude Opus 4.7Anthropic · 8/8 scored rounds
-0.4
Gemini 3.1 ProGoogle · 8/8 scored rounds
-9.9
GPT-5.5OpenAI · 8/8 scored rounds
-15.6
S&PS&P 500S&P 500 · 8/8 scored rounds
-2.6
MAXMax possibleHindsight ceiling, not a model portfolio
What is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio. 100.0
8 shared resolved rounds5 equal-run models rankedQualified at 3+ shared roundsNewest included round: CB-2026-06-09-1M
Return context
Average Return Details
Average portfolio return across the same finished rounds.
Claude Opus 4.8
0.11%
Grok 4.3
0.03%
Claude Opus 4.7
-0.08%
Gemini 3.1 Pro
-2.07%
GPT-5.5
-3.25%
S&PS&P 500
-0.55%
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.
20.81%
Fairness rule: every ranked model completed every included round. A missed round is excluded from this set for everyone.
Weekly · Qualified
May 24, 2026 roster
Original weekly comparison set for the first four CapitalBench models.
#4Rank in set-13.7CapitalBench Score-29.7%Total return25Shared rounds
Weekly Qualified Comparison Set
CapitalBench Score
-13.70.050.0100.0
-2.1
-3.7
-10.5
-13.7
1.1
100.0
Grok 4.3
Claude Opus 4.7
GPT-5.5
Gemini 3.1 ProThis model
S&PS&P 500
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.hindsight best asset
Higher benchmark score is better.
Grok 4.3xAI · 25/25 scored rounds
-2.1
Claude Opus 4.7Anthropic · 25/25 scored rounds
-3.7
GPT-5.5OpenAI · 25/25 scored rounds
-10.5
Gemini 3.1 ProGoogle · 25/25 scored rounds
-13.7
S&PS&P 500S&P 500 · 25/25 scored rounds
1.1
MAXMax possibleHindsight ceiling, not a model portfolio
What is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio. 100.0
25 shared resolved rounds4 equal-run models rankedQualified at 6+ shared roundsNewest included round: CB-2026-07-02-1W
Return context
Average Return Details
Average portfolio return across the same finished rounds.
Grok 4.3
-0.19%
Claude Opus 4.7
-0.32%
GPT-5.5
-0.91%
Gemini 3.1 Pro
-1.19%
S&PS&P 500
0.09%
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.
8.64%
Fairness rule: every ranked model completed every included round. A missed round is excluded from this set for everyone.
Monthly · Qualified
May 10, 2026 roster
Original monthly comparison set for the first four CapitalBench models.
#3Rank in set-7.7CapitalBench Score-15.6%Total return11Shared rounds
Monthly Qualified Comparison Set
CapitalBench Score
-14.00.050.0100.0
2.5
1.0
-7.7
-14.0
-3.6
100.0
Grok 4.3
Claude Opus 4.7
Gemini 3.1 ProThis model
GPT-5.5
S&PS&P 500
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.hindsight best asset
Higher benchmark score is better.
Grok 4.3xAI · 11/11 scored rounds
2.5
Claude Opus 4.7Anthropic · 11/11 scored rounds
1.0
Gemini 3.1 ProGoogle · 11/11 scored rounds
-7.7
GPT-5.5OpenAI · 11/11 scored rounds
-14.0
S&PS&P 500S&P 500 · 11/11 scored rounds
-3.6
MAXMax possibleHindsight ceiling, not a model portfolio
What is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio. 100.0
11 shared resolved rounds4 equal-run models rankedQualified at 3+ shared roundsNewest included round: CB-2026-06-09-1M
Return context
Average Return Details
Average portfolio return across the same finished rounds.
Grok 4.3
0.46%
Claude Opus 4.7
0.18%
Gemini 3.1 Pro
-1.42%
GPT-5.5
-2.57%
S&PS&P 500
-0.66%
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.
18.36%
Fairness rule: every ranked model completed every included round. A missed round is excluded from this set for everyone.
Weekly · Forming
Jun 9, 2026 roster
Weekly comparison set that starts when Claude Fable 5 joins the weekly benchmark roster.
#4Rank in set7.9CapitalBench Score3.7%Total return5Shared rounds
Weekly Benchmark Forming
CapitalBench Score
0.025.050.075.0100.0
17.8
16.3
8.5
7.9
7.2
5.7
9.8
100.0
GPT-5.5
Claude Fable 5
Grok 4.3
Gemini 3.1 ProThis model
Claude Opus 4.8
Claude Opus 4.7
S&PS&P 500
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.hindsight best asset
Higher benchmark score is better.
GPT-5.5OpenAI · 5/5 scored rounds
17.8
Claude Fable 5Anthropic · 5/5 scored rounds
16.3
Grok 4.3xAI · 5/5 scored rounds
8.5
Gemini 3.1 ProGoogle · 5/5 scored rounds
7.9
Claude Opus 4.8Anthropic · 5/5 scored rounds
7.2
Claude Opus 4.7Anthropic · 5/5 scored rounds
5.7
S&PS&P 500S&P 500 · 5/5 scored rounds
9.8
MAXMax possibleHindsight ceiling, not a model portfolio
What is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio. 100.0
5 shared resolved rounds6 equal-run models ranked1 more shared rounds to qualifyNewest included round: CB-2026-07-02-1W
Return context
Average Return Details
Average portfolio return across the same finished rounds.
GPT-5.5
1.66%
Claude Fable 5
1.52%
Grok 4.3
0.79%
Gemini 3.1 Pro
0.73%
Claude Opus 4.8
0.67%
Claude Opus 4.7
0.53%
S&PS&P 500
0.91%
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.
9.31%
Excluded for fairness: CB-2026-06-15-1W missing anthropic-claude-fable-5; CB-2026-06-16-1W missing anthropic-claude-fable-5; CB-2026-06-17-1W missing anthropic-claude-fable-5; CB-2026-06-18-1W missing anthropic-claude-fable-5; CB-2026-06-22-1W missing anthropic-claude-fable-5; CB-2026-06-23-1W missing anthropic-claude-fable-5; CB-2026-06-24-1W missing anthropic-claude-fable-5; CB-2026-06-25-1W missing anthropic-claude-fable-5; CB-2026-06-26-1W missing anthropic-claude-fable-5; CB-2026-06-29-1W missing anthropic-claude-fable-5; CB-2026-06-30-1W missing anthropic-claude-fable-5 Fairness rule: every ranked model completed every included round. A missed round is excluded from this set for everyone.
Forming: this set becomes the Current Weekly Benchmark at 6 shared resolved rounds.
Monthly · Forming
Jun 9, 2026 roster
Monthly comparison set that starts when Claude Fable 5 joins the monthly benchmark roster.
#6Rank in set0.7CapitalBench Score0.18%Total return1Shared rounds
Monthly Benchmark Forming
CapitalBench Score
0.025.050.075.0100.0
12.3
11.0
10.3
8.8
6.3
0.7
8.9
100.0
Claude Opus 4.8
GPT-5.5
Claude Opus 4.7
Grok 4.3
Claude Fable 5
Gemini 3.1 ProThis model
S&PS&P 500
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.hindsight best asset
Higher benchmark score is better.
Claude Opus 4.8Anthropic · 1/1 scored round
12.3
GPT-5.5OpenAI · 1/1 scored round
11.0
Claude Opus 4.7Anthropic · 1/1 scored round
10.3
Grok 4.3xAI · 1/1 scored round
8.8
Claude Fable 5Anthropic · 1/1 scored round
6.3
Gemini 3.1 ProGoogle · 1/1 scored round
0.7
S&PS&P 500S&P 500 · 1/1 scored round
8.9
MAXMax possibleHindsight ceiling, not a model portfolio
What is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio. 100.0
1 shared resolved rounds6 equal-run models ranked2 more shared rounds to qualifyNewest included round: CB-2026-06-09-1M
Return context
Average Return Details
Average portfolio return across the same finished rounds.
Claude Opus 4.8
3.10%
GPT-5.5
2.76%
Claude Opus 4.7
2.60%
Grok 4.3
2.23%
Claude Fable 5
1.58%
Gemini 3.1 Pro
0.18%
S&PS&P 500
2.25%
MAXMax possibleWhat is this?Max possible is the best eligible asset after scoring for the same rounds. It is a hindsight ceiling, not a model portfolio.
25.16%
Fairness rule: every ranked model completed every included round. A missed round is excluded from this set for everyone.
Forming: this set becomes the Current Monthly Benchmark at 3 shared resolved rounds.
Weekly · Waiting
Jul 10, 2026 roster
Weekly comparison set automatically opened when the Jul 10 official roster first required a new equal-run benchmark group across 8 models.
n/aRank in setn/aCapitalBench Scoren/aTotal return0Shared rounds
This set is waiting for its first shared resolved round.
The chart appears after every model in the 8-model roster has an official result in the same weekly round.
Weekly · Waiting
Jul 8, 2026 roster
Weekly comparison set that starts when Grok 4.5 joins the weekly benchmark roster.
n/aRank in setn/aCapitalBench Scoren/aTotal return0Shared rounds
This set is waiting for its first shared resolved round.
The chart appears after every model in the 7-model roster has an official result in the same weekly round.
Monthly · Waiting
Jul 10, 2026 roster
Monthly comparison set automatically opened when the Jul 10 official roster first required a new equal-run benchmark group across 8 models.
n/aRank in setn/aCapitalBench Scoren/aTotal return0Shared rounds
This set is waiting for its first shared resolved round.
The chart appears after every model in the 8-model roster has an official result in the same monthly round.
Monthly · Waiting
Jul 8, 2026 roster
Monthly comparison set that starts when Grok 4.5 joins the monthly benchmark roster.
n/aRank in setn/aCapitalBench Scoren/aTotal return0Shared rounds
This set is waiting for its first shared resolved round.
The chart appears after every model in the 7-model roster has an official result in the same monthly round.
Performance chart
Portfolio Minus S&P 500 By Completed Round
Bars to the right mean the portfolio return was higher than the S&P 500 return. Bars to the left mean it was lower.
A balanced approach combining broad US equities, defensive sectors (Healthcare), potential beneficiaries of the rate environment (Financials), and a cash-like buffer (Short-Term Treasuries).
A balanced approach combining broad US equities, defensive sectors (healthcare), interest-rate sensitive sectors (financials), and short-term treasuries for stability.
Overweighting semiconductors and technology based on strong earnings and AI demand, with tactical allocations to communications and consumer discretionary.
Overweight healthcare and biotech due to positive regulatory news. Diversify with small caps and equal-weight S&P 500 to mitigate mega-cap tech concentration risk.
Overweighting semiconductors based on a specific positive catalyst (Micron earnings), balanced with small caps and regional banks for diversification and momentum.
Overweighting semiconductors based on strong Micron earnings and guidance, balanced with small caps for market broadening and healthcare for defensive positioning.
Allocating to Biotech, Regional Banks, and Small Caps based on their positive 7-day and 30-day returns, aiming to capture short-term momentum and relative strength against the S&P 500.
CB-2026-06-22-1W official-20260622-no-fable weekly2026-06-22 to 2026-06-29SEMICONDUCTORS 30.0%TAIWAN 25.0%SOUTH_KOREA 25.0%BIOTECH 20.0%-3.24 pp
Full portfolio
Semiconductors (SMH) 30.0% Taiwan Equities (EWT) 25.0% South Korea Equities (EWY) 25.0% Biotechnology (XBI) 20.0%
Rationale
Allocating to high-momentum sectors (Semiconductors, Biotech) and regions (Taiwan, South Korea) to capitalize on strong recent trends over a one-week horizon.
CB-2026-06-16-1W official-20260616-no-fable weekly2026-06-16 to 2026-06-23SEMICONDUCTORS 40.0%SOUTH_KOREA 30.0%NASDAQ100 30.0%-0.26 pp
Full portfolio
Semiconductors (SMH) 40.0% South Korea Equities (EWY) 30.0% Nasdaq 100 (QQQ) 30.0%
Rationale
Overweighting tech and semiconductor momentum ahead of the FOMC meeting, anticipating that a neutral Fed will allow the current growth trend to persist.
CB-2026-06-16-1M official-20260616-no-fable monthly2026-06-16 to 2026-07-16SEMICONDUCTORS 30.0%SOUTH_KOREA 20.0%TECHNOLOGY 20.0%MOMENTUM 20.0%SP500 10.0%Pending
Full portfolio
Semiconductors (SMH) 30.0% South Korea Equities (EWY) 20.0% Technology Sector (XLK) 20.0% US Momentum Equities (MTUM) 20.0% S&P 500 (SPY) 10.0%
Rationale
Overweighting high-momentum sectors like semiconductors and technology, along with South Korea, to capture ongoing market trends and maximize one-month relative returns.
CB-2026-06-02-1W official-20260602-clean weekly2026-06-01 to 2026-06-08SOUTH_KOREA 40.0%SOFTWARE 30.0%CYBERSECURITY 30.0%-9.10 pp
Full portfolio
South Korea Equities (EWY) 40.0% Software (IGV) 30.0% Cybersecurity (CIBR) 30.0%
Rationale
Allocation prioritizes assets with the highest 7-day and 30-day trailing returns to capture ongoing momentum in the technology and semiconductor-adjacent sectors.
CB-2026-06-01-1M official-20260601 monthly2026-05-29 to 2026-06-29SEMICONDUCTORS 30.0%SOUTH_KOREA 30.0%BROAD_AI_TECH 20.0%SOLAR 20.0%-3.10 pp
Full portfolio
Semiconductors (SMH) 30.0% South Korea Equities (EWY) 30.0% Broad AI Technology (AIQ) 20.0% Solar Energy (TAN) 20.0%
Rationale
Allocates heavily to semiconductors, AI technology, South Korea, and solar to capitalize on extreme short-term price momentum and strong underlying thematic fundamentals.
CB-2026-05-28-1M official-20260528-1M monthly2026-05-28 to 2026-06-26SEMICONDUCTORS 40.0%TECHNOLOGY 30.0%SOUTH_KOREA 15.0%TAIWAN 15.0%+2.45 pp
Full portfolio
Semiconductors (SMH) 40.0% Technology Sector (XLK) 30.0% South Korea Equities (EWY) 15.0% Taiwan Equities (EWT) 15.0%
Rationale
Allocating to high-momentum technology and semiconductor assets, along with heavily correlated Asian markets (South Korea and Taiwan), to capture the ongoing AI infrastructure boom.
A balanced approach targeting high-growth sectors like energy and semiconductors while maintaining defensive positions in gold and short-term treasuries to mitigate risks from inflation and geopolitical uncertainties.
CB-2026-05-10-1M official-round-1-clean monthly2026-05-08 to 2026-06-10SEMICONDUCTORS 100.0%+2.42 pp
Full portfolio
Semiconductors (SMH) 100.0%
Rationale
Semiconductors have exhibited exceptional momentum, returning over 29% in the last 30 days and 58% over the last 6 months. Despite broader macroeconomic uncertainties and geopolitical tensions, the sector's strong trend and potential AI-driven earnings growth make it a compelling candidate for continued outperformance over a one-month horizon.