SmartPlay
Language-model agents play a collection of games designed to test planning, reasoning, memory and interaction abilities.
The official repository has 145 GitHub stars and supports multiple games and agent evaluation protocols.
- Made by
- Microsoft Research
- Released
- October 2023
- Kind
- Puzzles and chess
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- A simulator
- Items from
- Generated
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-4 Mar 2023 #181 on the capability index
- Last model added
- July 2023
- Scores refreshed
- November 21st, 2026
GPT-4 leads with 0.80 among models released by July 2023 GPT-4 leads with 0.80 among models released by July 2023
Best score per entry from the benchmark's paper, read on November 21st, 2026. A paper’s table doesn’t add newer models. The paper reports no overall score, so this is the plain mean of its seven game scores. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in July 2023, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 GPT-4 Mar 2023 0.80
- 2 GPT-4 Jun 2023 0.78
- 3 text-davinci-003 0.51
- 4 Claude (snapshot unspecified) 0.49
- 5 Bard 0.49
- 6
LLaMA-13B 0.32
- 7
Llama 2-13B 0.30
- 8 vicuna-13b 0.21
Source: The SmartPlay paper
- 1 GPT-4 Mar 2023 0.80
- 2 GPT-4 Jun 2023 0.78
- 3
LLaMA-13B 0.32
- 4
Llama 2-13B 0.30
Source: The SmartPlay paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-4 Mar 2023 | 0.80 | Mar 2023 | Official leaderboard | |
| 2 | GPT-4 Jun 2023 | 0.78 | Jun 2023 | Official leaderboard | |
| 3 | text-davinci-003 | 0.51 | Official leaderboard | ||
| 4 | Claude (snapshot unspecified) | 0.49 | Official leaderboard | ||
| 5 | Bard | 0.49 | Official leaderboard | ||
| 6 | | 0.32 | Feb 2023 | Official leaderboard | |
| 7 | | 0.30 | Jul 2023 | Official leaderboard | |
| 8 | vicuna-13b | 0.21 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-4 Mar 2023 | 0.80 | Mar 2023 | Official leaderboard | |
| 2 | GPT-4 Jun 2023 | 0.78 | Jun 2023 | Official leaderboard | |
| 3 | | 0.32 | Feb 2023 | Official leaderboard | |
| 4 | | 0.30 | Jul 2023 | Official leaderboard |
Rankings by part
We read 7 parts from the benchmark’s paper, each ranking models on one part of the test, such as bandit, rps and hanoi. GPT-4 leads 4 of them, and GPT-4 leads the other 3. Arrows show how far a model moved from its overall rank among the models we track.
bandit
4 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 1.00
- 2 GPT-4 Mar 2023 ↓1 0.97
- 3
LLaMA-13B 0.68
- 4
Llama 2-13B 0.50
rps
4 models we track have a score here.
- 1 GPT-4 Mar 2023 0.98
- 2 GPT-4 Jun 2023 0.91
- 3
LLaMA-13B 0.50
- 4
Llama 2-13B 0.35
hanoi
4 models we track have a score here.
- 1 GPT-4 Mar 2023 0.90
- 2 GPT-4 Jun 2023 0.83
- 3
Llama 2-13B ↑1 0.37
- 4
LLaMA-13B ↓1 0.33
messengerL1
4 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 0.90
- 2 GPT-4 Mar 2023 ↓1 0.87
- 3
LLaMA-13B 0.16
- 4
Llama 2-13B 0.12
messengerL2
4 models we track have a score here.
- 1 GPT-4 Mar 2023 0.97
- 2 GPT-4 Jun 2023 0.93
- 3
Llama 2-13B ↑1 0.13
- 4
LLaMA-13B ↓1 0.06
crafter
4 models we track have a score here.
- 1 GPT-4 Mar 2023 0.32
- 2 GPT-4 Jun 2023 0.26
- 3
LLaMA-13B 0.04
- 4
Llama 2-13B 0.04
minecraft
4 models we track have a score here.
- 1 GPT-4 Jun 2023 ↑1 0.61
- 2
Llama 2-13B ↑2 0.61
- 3 GPT-4 Mar 2023 ↓2 0.59
- 4
LLaMA-13B ↓1 0.50
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The SmartPlay paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official SmartPlay leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Games and long tasks
- ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6.2 Astra 100%
- Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games.
Gemini 2.5 Pro 3.90
- MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks.
Seed2.0 Pro 49%
- TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress.
Llama 4 Scout 7.00
- GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
- Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 3 years ago, in July 2023, for Llama 2-13B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (8 entries).
- Pass Models scored 4 models we track, plus 4 other entries on its leaderboard.
- Warning Newest score The newest came 3 years ago, in July 2023, for Llama 2-13B.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-4 (#181).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 4 rows for models we track and 4 other entries at the last read.
- Pass Last read November 21st, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.