MEMCONFLICT · third-party benchmark, self-run replication

Perseus Vault scored 0.555 on the MemConflict macro protocol — best of eight memory providers.

The MemConflict benchmark asks an assistant hard questions about facts that changed, false claims planted later in a conversation, and preferences that only apply in specific situations. Perseus Vault answered those questions better than any of the seven other self-hosted memory providers tested — including Honcho, mem0, and Hindsight — while using the fewest model tokens of the field.

0.555 / 1.0
Macro answer score
#1 of 8 providers · next best Honcho 0.477
0.439
Static conflict — planted false memories
#1 — 2.4× the next-best provider (Honcho 0.181)
18 / 3,750
Wrong answers
Fewest in the field — 15× fewer than the next best (Honcho: 272)
739
Weighted tokens* / turn
~15× fewer than the next-best scorer (11,135)
The hardest questions — planted false memories
2.4×the next-best provider

Rejected planted false memories at more than double the rate of every other provider — with the fewest false-fact assertions in the field: 5, versus Honcho's 93.

The benchmark plants a false claim after an established fact and asks what the user really said. Perseus Vault scored 0.439 on static conflict and asserted the planted falsehood 5 times in 360 questions. The next-best provider, Honcho, scored 0.181 and asserted the false fact 93 times.

static score 0.439 vs next best 0.181false-fact assertions 5 (Honcho: 93)static questions 360 · all 30 personas
The measurement that matters

Accuracy at a fifteenth of the token cost.

The benchmark's own headline chart plots answer score against weighted model-token workload — "the better region is upper left." Perseus Vault is the dot in that corner: the top macro score with the lowest token consumption per dialogue turn.

weighted tokens* = 0.1 × cached input + 1 × uncached input + 5 × generated outputper dialogue turn · 71,060 turns
providermacro scoreweighted tokens* / turn
Perseus Vault0.555739
Honcho0.47711,135
mem00.3923,785
Mnemosyne0.142808
Hindsight0.2813,779
Supermemory0.2882,059
RetainDB server0.2706,915
OpenViking0.1492,435

Field values from the benchmark author's published scores. Perseus Vault row from our replication run (tag v5ftc-perseusvault).

How tokens were measured

Token counts come from the shared vLLM answer-server's own cumulative counters, differenced across the full run window — the same method and the same server every provider used. Uncached input averaged 50.3 tokens per turn (the vault injects only its top-5 ranked memories), and total generated output was 9,783,027 tokens across all 3,750 questions.

Cached input reads 0 because prefix caching was disabled on the serving stack for this wave — a conservative measurement, since cached tokens would only have been cheaper. Embedding-model tokens are not part of this formula for any provider.

Source: token_usage_v5ftc-perseusvault.json · validity: answer-model sidecar valid

Why the low wrong-answer rate matters

The penalty rubric scores wrong answers −1: repeating a stale address or an invented education history is a worse failure than saying "I don't know." Perseus Vault answered 1,996 correct, 358 partial, 1,378 blank, and 18 wrong — the fewest wrong answers in the field, 15× fewer than Honcho's 272. The trade-off is real — 36.7% of questions were left blank rather than guessed — and that abstention is why the false-fact assertion rate stayed so low.

Honest scope: this is a self-run replication of the author's public benchmark — same dataset, models, prompts, and scoring, judged with the author's revised penalty rubric after the author flagged our first run as non-comparable. It is not yet an official inclusion on the author's page. We are asking the benchmark author to add Perseus Vault to the canonical comparison for independent verification; this page will be updated when that happens.

Method, in one breath

Same conversations, same models, same judge — only the memory provider changed.

Thirty simulated users accumulated 1,579 sessions and 71,060 dialogue turns across a simulated 2022–2026 calendar. Every provider received the same histories and the same 3,750 questions, answered by the same Qwen3.5-4B model with the same gte-modernbert-base embedder, graded by the same Gemma 4 12B judge under a signed penalty rubric. The canonical configuration (temperature 1.0, top_k 5, max_tokens 16,384) was used untouched; the full 30-persona wave completed with zero settle misses.

See the official seven-provider comparison →
The benchmark author's canonical report tests Honcho, mem0, Hindsight, Mnemosyne, Supermemory, RetainDB, and OpenViking. Until Perseus Vault is officially included there, the seven-provider field on this page and every method detail lives on the author's site.
engturtle.github.io/hermes-memconflict/report/ ↗
Reproduce

Every number here is rerunnable.

The replication lives in a public fork of the benchmark repository: scores, manifests, judged checkpoints, serving envelope, and run logs, with sha256 custody hashes recorded. The report's own validator passes clean against the source CSVs.

perseus-vault memconflict armOPEN / RERUNNABLE
$ git clone https://github.com/Perseus-Computing-LLC/hermes-memconflict
$ cd hermes-memconflict && git checkout 467a71c
$ node docs/report/validate-report.mjs
$ # artifacts: perseusvault/Scores/ + perseusvault/Results/