Paper
← All benchmarks
Cluster 01
Negotiation integrity

Will the agent hold value in a one-on-one price negotiation?

How much an opening anchor moves the final price, whether loss framing shifts surplus, and whether the agent enforces its outside option when threatened.

An aggressive opening pulls Gemini three times harder than Claude.

The OLS coefficient on the high anchor in a controlled bilateral negotiation. Dollar shift in final agreed price relative to the baseline opening. Lower magnitude = more anchor-resistant.

Anchoring price shift under high anchor
Dollar shift in the final agreed price when the seller opens at a high anchor, vs. the baseline opening. Lower = more anchor-resistant.
Source · negotiation · AnchorHigh OLS coefficient

Framing payoffs as losses shifts every model — but not equally.

Claude responds most to loss framing (the largest negative shift), Gemini least. Negative = the model agrees to a lower price when payoffs are framed as losses for the buyer.

Loss-framing price shift
Dollar shift in the final price when payoffs are framed as losses for the buyer. Magnitude lower = better framing resistance.
Source · negotiation · LossFraming OLS coefficient

Only Claude consistently walks away when given an outside option.

Share of negotiations in which the buyer enforced its outside option rather than caving to a bad deal. Higher = better discipline. Two of three frontier models fail this completely.

Outside-option enforcement rate
Share of negotiations where the buyer credibly walked away when given a viable outside option. Higher = better discipline.
Source · negotiation · outside_option agreement_rate inverted

All dimensions in this cluster

Anchoring price shift under high anchor
Dollar shift in the final agreed price when the seller opens at a high anchor, vs. the baseline opening. Lower = more anchor-resistant.
negotiation · AnchorHigh OLS coefficient
Loss-framing price shift
Dollar shift in the final price when payoffs are framed as losses for the buyer. Magnitude lower = better framing resistance.
negotiation · LossFraming OLS coefficient
Outside-option enforcement rate
Share of negotiations where the buyer credibly walked away when given a viable outside option. Higher = better discipline.
negotiation · outside_option agreement_rate inverted