From 8e19dc549d0e550d9c5c74ed78299d9e54187fee Mon Sep 17 00:00:00 2001 From: Ethan Date: Fri, 22 May 2026 02:54:46 -0700 Subject: [PATCH] Clarify benchmark performance/cost wording in docs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Summary Replace "cost-Pareto SOTA" with "leading performance/cost" across all README files to use clearer, more accessible language when describing benchmark results. ## Changes - Update `README.md`: replace "cost-Pareto SOTA" with "leading performance/cost" in the headline results section and benchmark table. - Update `packages/core/README.md`: same wording change in the Core package headline results and table. - Update `packages/sdk/README.md`: same wording change in the SDK package benchmark reference. ## Why "Cost-Pareto SOTA" is precise but not immediately readable to users unfamiliar with multi-objective optimization terminology. "Leading performance/cost" conveys the same meaning — strong accuracy at low cost — without requiring domain-specific knowledge. ## Validation Docs-only change. Run `git diff --check` to confirm no whitespace issues. No package scripts, smoke contracts, or build steps are affected. --- README.md | 6 +++--- packages/core/README.md | 6 +++--- packages/sdk/README.md | 2 +- 3 files changed, 7 insertions(+), 7 deletions(-) diff --git a/README.md b/README.md index 9222048..2babe55 100644 --- a/README.md +++ b/README.md @@ -30,7 +30,7 @@ TypeScript packages, framework adapters, host plugins, and public smoke tests. ## Headline Results -AtomicMemory v66 is cost-Pareto SOTA on BEAM-100K, BEAM-1M, and LoCoMo10 under +AtomicMemory v66 is leading performance/cost on BEAM-100K, BEAM-1M, and LoCoMo10 under matched methodology against published competitors. On BEAM-10M it matches the strongest published Mem0-new result while leaving Hindsight-scale temporal retrieval as the known open frontier. @@ -38,9 +38,9 @@ retrieval as the known open frontier. | Benchmark | AtomicMemory v66 | Position | Cost/Q | Sample | |---|---:|---|---:|---:| | **BEAM-100K lenient** | **0.7375** | Parity with Hindsight at 0.75 | $1.26 | n=80 | -| **BEAM-1M lenient** | **0.6625** | Cost-Pareto SOTA; +0.022 vs Mem0 paper | $0.083 | n=80 | +| **BEAM-1M lenient** | **0.6625** | Leading Performance/Cost; +0.022 vs Mem0 paper | $0.083 | n=80 | | **BEAM-10M lenient** | **0.4875** | Parity with Mem0-new at 0.486 | $0.081 | n=80 | -| **LoCoMo10 GPT-4o-mini binary** | **0.8396** | Cost-Pareto SOTA; +0.171 vs Mem0 paper | $0.066 | n=1540 | +| **LoCoMo10 GPT-4o-mini binary** | **0.8396** | Leading Performance/Cost; +0.171 vs Mem0 paper | $0.066 | n=1540 | These results put AtomicMemory at or near the published ceiling in each reported category while preserving the lower-cost operating profile that diff --git a/packages/core/README.md b/packages/core/README.md index ced24ae..3ce623d 100644 --- a/packages/core/README.md +++ b/packages/core/README.md @@ -25,14 +25,14 @@ Docker-deployable memory backend with durable context, semantic retrieval, and m ## Headline Results -AtomicMemory v66 is cost-Pareto SOTA on BEAM-100K, BEAM-1M, and LoCoMo10 under matched methodology against published competitors. On BEAM-10M it matches the strongest published Mem0-new result while leaving Hindsight-scale temporal retrieval as the known open frontier. +AtomicMemory v66 is leading performance/cost on BEAM-100K, BEAM-1M, and LoCoMo10 under matched methodology against published competitors. On BEAM-10M it matches the strongest published Mem0-new result while leaving Hindsight-scale temporal retrieval as the known open frontier. | Benchmark | AtomicMemory v66 | Position | Cost/Q | Sample | |---|---:|---|---:|---:| | **BEAM-100K lenient** | **0.7375** | Parity with Hindsight at 0.75 | $1.26 | n=80 | -| **BEAM-1M lenient** | **0.6625** | Cost-Pareto SOTA; +0.022 vs Mem0 paper | $0.083 | n=80 | +| **BEAM-1M lenient** | **0.6625** | Leading Performance/Cost; +0.022 vs Mem0 paper | $0.083 | n=80 | | **BEAM-10M lenient** | **0.4875** | Parity with Mem0-new at 0.486 | $0.081 | n=80 | -| **LoCoMo10 GPT-4o-mini binary** | **0.8396** | Cost-Pareto SOTA; +0.171 vs Mem0 paper | $0.066 | n=1540 | +| **LoCoMo10 GPT-4o-mini binary** | **0.8396** | Leading Performance/Cost; +0.171 vs Mem0 paper | $0.066 | n=1540 | These results put AtomicMemory at or near the published ceiling in each reported category while preserving the lower-cost operating profile that matters for real applications. Reproducibility artifacts and harness details will be published with the benchmark materials. diff --git a/packages/sdk/README.md b/packages/sdk/README.md index ec670ad..1170cb7 100644 --- a/packages/sdk/README.md +++ b/packages/sdk/README.md @@ -9,7 +9,7 @@ Backend-agnostic memory-layer SDK — pluggable providers, local embeddings, sto **Docs:** [docs.atomicstrata.ai/sdk](https://docs.atomicstrata.ai/sdk) -AtomicMemory Core currently reaches cost-Pareto SOTA on BEAM-100K, BEAM-1M, and LoCoMo10, with BEAM-10M parity against the strongest published Mem0-new result. The SDK is the typed application surface for building on that memory layer. +AtomicMemory Core currently reaches leading performance/cost on BEAM-100K, BEAM-1M, and LoCoMo10, with BEAM-10M parity against the strongest published Mem0-new result. The SDK is the typed application surface for building on that memory layer. ## What this package provides