JAM with a Qwen3.5-4B Researcher scores higher than the evaluated training-free memory systems and larger trained agents across most main benchmark settings; the quoted Memory-R1 variants have higher open-domain LoCoMo F1. On HotpotQA it sustains comparatively strong performance as context grows from 56k to 224k, although its F1 decreases with length.

Reported and observed measurements

Assessments (0)

No immutable Assessment has been published for this Claim yet.

Runs (0)

No execution has been linked to this Claim yet.

Each row is a measurement with its reported value, latest available observation, and cumulative evidence state.
Metric and conditions	Reported	Latest observation	Evidence judgment
method: VANILLA · baseLLM: Qwen3.5-4B · category: SH · benchmark: locomo	48.15 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: MH · benchmark: locomo	32.98 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: TE · benchmark: locomo	42.66 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: OD · benchmark: locomo	19.63 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: OA · benchmark: locomo	42.45 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: LME · benchmark: longmemeval	54.8%	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: NAQA · benchmark: narrativeqa	31.23 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: 56k · benchmark: hotpotqa-56k	63.56 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: 112k · benchmark: hotpotqa-112k	53.04 F1 points (0–100)	—	Not assessed
method: VANILLA · baseLLM: Qwen3.5-4B · category: 224k · benchmark: hotpotqa-224k	36.97 F1 points (0–100)	—	Not assessed
method: RAG · baseLLM: Qwen3.5-4B · category: SH · benchmark: locomo	49.04 F1 points (0–100)	—	Not assessed
method: RAG · baseLLM: Qwen3.5-4B · category: MH · benchmark: locomo	34.39 F1 points (0–100)	—	Not assessed

Metric and conditions

Reported

Latest observation

Evidence judgment

method: VANILLA · baseLLM: Qwen3.5-4B · category: SH · benchmark: locomo

48.15 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: MH · benchmark: locomo

32.98 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: TE · benchmark: locomo

42.66 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: OD · benchmark: locomo

19.63 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: OA · benchmark: locomo

42.45 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: LME · benchmark: longmemeval

54.8%

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: NAQA · benchmark: narrativeqa

31.23 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: 56k · benchmark: hotpotqa-56k

63.56 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: 112k · benchmark: hotpotqa-112k

53.04 F1 points (0–100)

—

Not assessed

method: VANILLA · baseLLM: Qwen3.5-4B · category: 224k · benchmark: hotpotqa-224k

36.97 F1 points (0–100)

—

Not assessed

method: RAG · baseLLM: Qwen3.5-4B · category: SH · benchmark: locomo

49.04 F1 points (0–100)

—

Not assessed

method: RAG · baseLLM: Qwen3.5-4B · category: MH · benchmark: locomo

34.39 F1 points (0–100)

—

Not assessed

Reported and observed measurements

Assessments (0)

Runs (0)

Measurement results