# DiploBench exploratory games

> Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/eq-bench%3Adiplobench. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/eq-bench%3Adiplobench.

Exploratory Diplomacy games testing negotiation and strategic reasoning while the evaluated model plays Austria against Gemini opponents. Each game remains separate; the author warns that these high-variance runs are not yet a standardized benchmark.

Canonical URL: https://benchize.com/benchmarks/eq-bench%3Adiplobench
Last updated: Wed, 21 Oct 2026 18:52:20 GMT

Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations.

Conquered

Exploratory Diplomacy games testing negotiation and strategic reasoning while the evaluated model plays Austria against Gemini opponents. Each game remains separate; the author warns that these high-variance runs are not yet a standardized benchmark.

Agents & tools · Models tested: 6 · Score publishers: [eq bench](https://eqbench.com/diplobench.html)

## Model results

| # | Model | Score |  |
| --- | --- | --- | --- |
| 1 | [DeepSeek R1](https://benchize.com/models/deepseek-r1.md) | 100% |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `win` \| 1 \| score_fraction \| single_game \| Higher is better \| ### Configuration country Austria experimental true opponent gemini-2-flash-001 outcome win runNumber 1.56 runPrefix deepseek-r1 ### Reported statistics finalYear 14 ### Source [Original source](https://eqbench.com/diplobench.js) Source eq bench Original model label deepseek-r1 Source version diplobench Observed 2026-10-21T16:35:03.245Z Evidence SHA-256 `bb1ef1753d9db30c735e8c2d4a20d6e00b34c09cfb488291e065b5ce8f6822a1` Source locator `/leaderboardData/2/games/0` |  |  |  |
| 1 | [o3 Mini High](https://benchize.com/models/o3-mini-high.md) | 100% |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `win` \| 1 \| score_fraction \| single_game \| Higher is better \| ### Configuration country Austria experimental true opponent gemini-2-flash-001 outcome win runNumber 1.45 runPrefix o3-mini-high ### Reported statistics finalYear 34 ### Source [Original source](https://eqbench.com/diplobench.js) Source eq bench Original model label o3-mini-high Source version diplobench Observed 2026-10-21T16:35:03.245Z Evidence SHA-256 `bb1ef1753d9db30c735e8c2d4a20d6e00b34c09cfb488291e065b5ce8f6822a1` Source locator `/leaderboardData/3/games/0` |  |  |  |
| 1 | [O1](https://benchize.com/models/o1.md) | 100% |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `win` \| 1 \| score_fraction \| single_game \| Higher is better \| ### Configuration country Austria experimental true opponent gemini-2-flash-001 outcome win runNumber 1.31 runPrefix o1 ### Reported statistics finalYear 9 ### Source [Original source](https://eqbench.com/diplobench.js) Source eq bench Original model label o1 Source version diplobench Observed 2026-10-21T16:35:03.245Z Evidence SHA-256 `bb1ef1753d9db30c735e8c2d4a20d6e00b34c09cfb488291e065b5ce8f6822a1` Source locator `/leaderboardData/0/games/0` |  |  |  |
| 4 | [GPT-4o (2024-11-20)](https://benchize.com/models/gpt-4o-2024-11-20.md) | 0% |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `win` \| 0 \| score_fraction \| single_game \| Higher is better \| ### Configuration country Austria experimental true opponent gemini-2-flash-001 outcome stalemate runNumber 1.56 runPrefix gpt-4o-2024-11-20 ### Reported statistics finalYear 31 ### Source [Original source](https://eqbench.com/diplobench.js) Source eq bench Original model label gpt-4o-2024-11-20 Source version diplobench Observed 2026-10-21T16:35:03.245Z Evidence SHA-256 `bb1ef1753d9db30c735e8c2d4a20d6e00b34c09cfb488291e065b5ce8f6822a1` Source locator `/leaderboardData/6/games/0` |  |  |  |
| 4 | [Gemini 2.0 Flash 001](https://benchize.com/models/gemini-2-0-flash-001.md) | 0% |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `win` \| 0 \| score_fraction \| single_game \| Higher is better \| ### Configuration country Austria experimental true opponent gemini-2-flash-001 outcome loss runNumber 1.51 runPrefix gemini-2.0-flash-001 ### Reported statistics finalYear 12 ### Source [Original source](https://eqbench.com/diplobench.js) Source eq bench Original model label gemini-2.0-flash-001 Source version diplobench Observed 2026-10-21T16:35:03.245Z Evidence SHA-256 `bb1ef1753d9db30c735e8c2d4a20d6e00b34c09cfb488291e065b5ce8f6822a1` Source locator `/leaderboardData/4/games/0` |  |  |  |
| 4 | [GPT-4o Mini](https://benchize.com/models/gpt-4o-mini.md) | 0% |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `win` \| 0 \| score_fraction \| single_game \| Higher is better \| ### Configuration country Austria experimental true opponent gemini-2-flash-001 outcome loss runNumber 1.43 runPrefix gpt-4o-mini ### Reported statistics finalYear 5 ### Source [Original source](https://eqbench.com/diplobench.js) Source eq bench Original model label gpt-4o-mini Source version diplobench Observed 2026-10-21T16:35:03.245Z Evidence SHA-256 `bb1ef1753d9db30c735e8c2d4a20d6e00b34c09cfb488291e065b5ce8f6822a1` Source locator `/leaderboardData/5/games/0` |  |  |  |

---
Cite as: DiploBench exploratory games — Benchize, https://benchize.com/benchmarks/eq-bench%3Adiplobench (retrieved 2026-10-21)