SmartPlay

https://github.com/Microsoft/SmartPlay

Microsoft Research's SmartPlay — LLM-agent benchmark on 6 games (Bandits, Rock-Paper-Scissors, Tower of Hanoi, MessengerEnv, Crafter, Minecraft). Tests 9 distinct capability axes including object understanding, planning, generalisation, error-handling, theory-of-mind. ICLR 2024. Bridges narrow-task benchmarks (MMLU) and full-agent benchmarks (AgentBench).

At a glance

Type
LLM-agent benchmark across 6 games — Microsoft Research
Tier
T3
Created
2023-10 (Wu et al. arXiv 2310.01557)
Latest release
not applicable — not OSS
License
MIT
Pricing
not applicable — open benchmark; free to use
Funding
not applicable — not commercial

Taxonomy

storage
n/a
retrieval
n/a
persistence
n/a
update
read-only
unit
n/a
governance
n/a
conflict
n/a

When to use

Optimised for: not applicable - eval dataset, not a system

Anti-fit: not applicable - eval dataset, not a system

Pros & cons

Pros

9 capability axes give finer-grained agent capability picture than single-task benchmarks; lightweight (LLM-interface, not full RL).

Cons

Text-only LLM interface bypasses perception; 6-game set is narrow; ICLR 2024 already feels dated vs newer agent benchmarks.

Claims & capabilities

ICLR 2024; 9 capability axes evaluated across 6 games; GPT-4 SOTA at time of release

Technical surface

API surface
not applicable — eval dataset, not a system
Backend storage
not applicable — eval dataset, not a system
Deployment
not applicable — run locally
Embedding model
not applicable — eval dataset, not a system
Multi-tenancy
not applicable — eval dataset, not a system
MCP
not applicable — benchmark / evaluation harness
A2A
not applicable — benchmark / evaluation harness
OpenTelemetry
not applicable — benchmark / evaluation harness

Similar systems

Other memory benchmarks & evaluation in the catalog, ranked by inbound references.

  • LoCoMo T3

    Conversations spanning up to 35 sessions / 300 turns / ~9K tokens, built from LLM-agent personas grounded on temporal event graphs and verified by human annotators. Tests QA (single-hop, multi-hop, temporal, commonsense, adversarial), event-graph summarisation, multimodal dialogue.

  • LongBench T3

    Tsinghua THUDM. 21 datasets across 6 categories (single-doc QA, multi-doc QA, summarisation, few-shot ICL, synthetic, code completion) in English + Chinese. ~6.7K words English / 13.4K chars Chinese average. v2 at ACL 2025 with harder, longer instances.

  • LongMemEval T3

    5 core memory abilities — info extraction, multi-session reasoning, temporal reasoning, knowledge updates, abstention — via 500 curated questions in synthetic chat histories. Two scales: LongMemEval-S (115K tokens), LongMemEval-M (up to 1.5M).

  • RULER T3

    NVIDIA. Extends NIAH with multi-needle retrieval, multi-hop entity tracing, aggregation, QA — configurable for arbitrary length / difficulty. Evaluated 17 long-context models on 13 tasks.

  • InfiniteBench (∞Bench) T3

    Tsinghua / OpenBMB. First benchmark with average input >100K tokens. 12 tasks across English + Chinese: long novel QA, code debugging, math in long context, summarisation, retrieval. Realistic documents (novels, code) rather than synthetic filler.

  • Atari 100k T3

    Sample-efficiency benchmark protocol on Arcade Learning Environment (ALE) — agents limited to 100k environment steps (~2 hours of game-play) before evaluation. Introduced by SimPLe (Kaiser et al. 2019) and adopted by EfficientZero, IRIS, BBF, DreamerV3, Storm. Tests world-model / model-based methods where memory-as-imagination matters more than huge replay buffers.

Row last verified 2026-05-14. Catalog data is CC-BY-4.0 — see how to read this.