Grok-2 in the lead after 5 earnings
So far Grok-2 up 36%, o1 +10%, Claude +9%, Gemini down 19%
The Flat Circle Benchmark measures the top language models’ ability to predict company earnings results. See our methodology for detail and disclaimers. If you haven’t already subscribed, join investors and engineers interested in LLMs+investment research here:
Key Takeaways
Grok currently leading with a +36% total return and a 60% hitrate across 5 Calls so far, Gemini lags with a 20% hitrate and total return of -19%.
The big driver for Grok and o1 were correctly calling SELL for Radius Recycling (RDUS), which declined 22% after earnings. Grok, o1 and Claude correctly anticipated continued challenging industry conditions, but did not anticipate a lack of insurance recoveries and an accounting adjustment which also contributed to results
All models have issued sell calls for Constellation Brands (STZ), Tilray (TLRY) and Walgreens (WBA)’s upcoming earnings
The system now has an improved methodology to identify peer companies, ensuring the models read a wider variety of peer ticker earnings transcripts that provide a sense of industry momentum going into earnings
Upcoming earnings
The complete reasoning for the above calls is available here. Please see our methodology for detail on how these Calls are generated.
Interesting takes
Tilray Brands, Inc. (TLRY)
Model: Google AI 2.0 Flash Experimental
Call: SELL
"...While they highlighted positive regulatory changes in Germany and Canada and strong growth in cannabis [last quarter], the market focused on the reduced guidance. This quarter highlights how important guidance is to the market and how it can override positive commentary... The company's increased investment in marketing and infrastructure in beverages creates a risk of lower profitability, while the company has yet to prove how they will succeed in the Delta-9 beverage space..."
Constellation Brands, Inc. (STZ)
Model: xAI grok-2-1212
Call: SELL
"The Beer business continues to perform well but faces macroeconomic headwinds... The recent downgrade by TD Cowen and other analysts' concerns about back-half performance suggest that market sentiment is shifting. Additionally, the divestiture of SVEDKA... may not immediately improve the bottom line and could signal further challenges in the Wine & Spirits segment. Given the significant underperformance of the stock relative to the S&P (-13.32%) and the recent analyst downgrades, the market appears to be pricing in potential challenges..."
Delta Air Lines, Inc. (DAL)
Model: Google AI 2.0 Flash Experimental (Hitrate: 20%)
Call: BUY
"Despite this, analyst ratings remain positive... The company expects a strong Q4, with a 30% increase in earnings over last year... The company's focus on premium offerings and diversified revenue streams should continue to provide a tailwind... The industry is showing signs of rationalization... Barring any unforeseen short-term disruptions, I expect Delta to beat expectations, driving the share price higher."
WD-40 Company (WDFC)
Model: Google AI 2.0 Flash Experimental (Hitrate: 20%)
Call: BUY
“...The company has consistently delivered strong volume-driven sales growth for the last three quarters. Gross margins are improving and are now approaching the 55% target... The company also announced the divestiture of its home care business, which will likely be viewed positively... The strong performance of peer Oil-Dri provides additional confidence in the health of the sector. Although the company did not raise guidance, I think the market may be underestimating the ongoing momentum...”
Recent earnings
Reconciling pre-earnings calls vs company earnings
Radius Recycling (RDUS) - Reported Jan 7 (amc), closed down 23%
The key culprit in the miss was lower sequential volumes in ferrous, nonferrous, and finished steel that drove down operating leverage and led to the outsized loss.
xAI grok-2-1212 (SELL)
"The persistent challenges in scrap supply flows and the inability to fully expand metal spreads have been consistent across the past four quarters. The upcoming quarter's setup does not suggest a significant change in these dynamics.”OpenAI o1-preview (SELL)
“...challenging market conditions persist, including tight scrap supply and compressed margins…in prior quarters, even when operational improvements were reported, the stock price reacted negatively post-earnings. With no significant positive catalysts in the recent press releases and continued industry headwinds, it is likely that the upcoming earnings will not exceed market expectations sufficiently to drive the stock price up”Claude 3 Opus (SELL)
"The steel and metals recycling markets RDUS operates in continue to face headwinds from tight scrap flows and subdued demand. The CEO indicated in Q4 that they have not yet seen any loosening of supply flows. Without a meaningful improvement, RDUS will likely continue struggling to expand margins. There have been no positive catalysts from recent press releases or peer earnings to indicate an impending turnaround in Q1. The COO transition and board appointment are unlikely to impact near-term results".
RDUS also reported no insurance recoveries and a $1/ferrous ton detriment from inventory accounting, which the models did not anticipate.
Follow the progress of LLMs learning to make accurate earnings calls
To track the progress of LLMs in investment research, subscribe for free:
If you have any questions, suggestions or would like to participate in this project, reply to this email or reach out via X.