Mem

Memory systems, readers, and writers.

3 studies
6 comparisons
189 original files

All memory comparisons

Complete + bound counts complete answers with source-matched quotes. Each row is a separate test.

Memory comparison overview
Comparison	Complete + bound	Resources	Evidence scale
Readers · fixed-context replay → Luna / DeepSeek Flash 1017 Viewing	DeepSeek Flash 101711/13 Luna medium11/13	$0.004811–$0.0193694.54–5.05 s / request	13 inputs / profile2 profiles · one pass
Readers · fixed-context replay → Luna reasoning effort	Luna low10/13 Luna medium11/13 Luna none11/13	$0.004149–$0.0047273.63–4.95 s / request	13 inputs / profile3 profiles · one pass
Systems · native retrieval → Native memory systems	OpenViking11/13 RAGFlow rc111/13	$0.006922–$0.024261Time: withheld from comparison	13 inputs / profile2 profiles · one pass
Systems · nightly version check → Native memory systems	RAGFlow nightly11/13	Cost not reported4.74 s / request	13 inputs / profile1 profile · one pass
Writers · extraction replay → Luna / DeepSeek Flash 1017	Quality not measuredDeepSeek Flash 1017 · Luna medium	$0.001142–$0.0029144.57–8.31 s / request	3 inputs / profile2 profiles · one pass
Writers · extraction replay → Luna reasoning effort	Quality not measuredLuna low · Luna medium · Luna none	$0.000709–$0.0011203.52–7.72 s / request	4 inputs / profile3 profiles · one pass

Separate tests · no combined ranking.

说明

Costs are workload totals in USD; resource ranges span profiles within that row. Quote binding does not establish semantic entailment.

Readers · fixed-context replay

Luna / DeepSeek Flash 1017 · Fixed-context reader replay

↑ All comparisons

Refine comparison

Study

Test

13 inputs / profile · 2 profiles · October 20, 2026

Test method

Compare complete, source-bound answers first, then the resources used.

Results & evidence ↓

Complete + bound

Answers · higher is better

DeepSeek Flash 101711/13
Luna medium11/13

Mean time

Seconds / input · lower is better

DeepSeek Flash 10175.05 s
Luna medium4.54 s

Settled cost

USD / test · lower is better

DeepSeek Flash 1017$0.019369
Luna medium$0.004811

Results and evidence

Sort columns. Open Details for the profile and original record.

Expand a row for all metrics

Sort

Metric

Reader · 13 inputs. Values are specific to this test. Actual settled USD for the 13 reader requests, including reported cache effects. Whole non-streaming request time; one pass, not a stable latency ranking.
Profile version						Evidence	Details
Luna mediumgpt-6-luna	11/13	11/13	28/28	4.536 s	$0.004811	Result evidence Luna medium Reader · 13 inputs · 13 inputs per profile Profile Complete + bound answers Cost USD total DeepSeek Flash 1017 11/13 $0.019369 Luna mediumSelected 11/13 $0.004811 Complete + bound requires a complete answer and exact quote binding. Other test groups remain separate. All recorded measures Complete answers 11/13 Complete + bound answers 11/13 Quotes matched 28/28 Time s · mean 4.536 s Cost USD total $0.004811275 Configuration & original record Recorded date 2026-10-20 Model openai/gpt-6-luna Provider OpenAI Effort medium OutputAllowance 128,000 tokens Control Retained RAGFlow nightly calls Fixed-context reader replay Actual settled USD for the 13 reader requests, including reported cache effects. Whole non-streaming request time; one pass, not a stable latency ranking. One diagnostic pass with real provider calls. Sort order is descriptive. Source binding alone does not establish semantic support. Open original metrics.json ↗
DeepSeek Flash 1017deepseek-v4.1-flash-1017	11/13	11/13	35/35	5.051 s	$0.019369	Result evidence DeepSeek Flash 1017 Reader · 13 inputs · 13 inputs per profile Profile Complete + bound answers Cost USD total DeepSeek Flash 1017Selected 11/13 $0.019369 Luna medium 11/13 $0.004811 Complete + bound requires a complete answer and exact quote binding. Other test groups remain separate. All recorded measures Complete answers 11/13 Complete + bound answers 11/13 Quotes matched 35/35 Time s · mean 5.051 s Cost USD total $0.019369416 Configuration & original record Recorded date 2026-10-20 Model deepseek/deepseek-v4.1-flash-1017 Provider DeepSeek; no fallback Effort low OutputAllowance 393,216 tokens Fixed-context reader replay Actual settled USD for the 13 reader requests, including reported cache effects. Whole non-streaming request time; one pass, not a stable latency ranking. One diagnostic pass with real provider calls. Sort order is descriptive. Source binding alone does not establish semantic support. Open original metrics.json ↗

One pass · — Not measured.

说明

USD totals are rounded to 6 decimals. The value tooltip and original record retain the recorded precision. Sorting is descriptive.

Actual settled USD for the 13 reader requests, including reported cache effects. Whole non-streaming request time; one pass, not a stable latency ranking.

Recorded finding

Both readers: 11/13 complete + bound. DeepSeek settled reader spend was about 4× Luna spend on these retained inputs.

Protocol and metric definitions

Thirteen identical saved questions, prompts, and retrieved contexts. Completeness and exact quote binding remain separate checks.

Synthetic inputs · real provider calls · one pass · no repeated-trial interval.

Complete + bound: complete answer with every quote matched to its supplied source. Complete: complete answers, before quote binding. Quotes: matched / emitted quotes, with a separate denominator. Source binding alone does not establish semantic entailment.

Full conditions and limitations

Results