See how much smarter your AI could be by adding tools
We independently measure the tools AI agents use and publish what each one actually adds.
Agentic Memory Indexi
MethodologyAgentic Memory Index v0.1 · Aug 2026
Highlights
Accuracy, cost, and speed for every tool, measured the same way.
Agentic Memory Indexi
The overall score: how often each tool finds the right answer.
Costi
USD per 1,000 successful answers
Answer speedi
How long an answer takes end to end, same agent model on every tool.
Check your own setup
Your agent already knows which tools you run. Paste one prompt and it checks that setup against the measured data, then tells you in plain words what would actually help.
Which memory tool is right for your agents?
No tool wins everything: match each tool's strengths to the questions your agents actually ask.
Correct Answers by Task Typei
From direct recall to synthesis, each tool has its strengths.
Failure Attributioni
Know how each tool tends to fail before your agent depends on it.
Rankings and retention
Every tool's score, and whether what it stores survives.
Every tool's score, cost, and speed in one table.
| # | Service | Index ↓ | Cost / 1k | Lookups / sessionLookups |
|---|---|---|---|---|
| 1 |  Karpathy Wiki |
98.5 | $568.93 | 1.6 |
| 2 |  Mitosis Cortex |
96.9 | $418.40 | 5.1 |
| 3 |  gbrain |
92.9 | $471.40 | 5.3 |
| 4 |  Hyperspell |
92.4 | $429.71 | 6.8 |
| 5 |  Mem0 |
92.3 | $341.42 | 5.1 |
| 6 |  Anthropic Memory |
84.0 | $419.45 | 8.4 |
| 7 |  Supermemory |
77.1 | $519.04 | 5.7 |
| 8 |  Zep |
75.1 | $394.75 | 5.1 |
| — | Claude Code built-in memoryClaude Code memory | 67.7 | $426.90 | — |
Long-term retentioni
How well early facts survive: probed at five checkpoints across 56 working sessions.
Agentic Search Index
Nine web search tools measured on a shared bank of live research questions.
Speed
One-time setup, indexing lag, and per-operation latency.
Onboarding, readiness, and store-session times for every tool.
| Tool | Onboard ↑ | Time to ready | Time to store |
|---|---|---|---|
| Anthropic Memory |
0.8 min | local | 56.0s |
| Karpathy Wiki |
1.0 min | local | 62.0s |
| Supermemory |
1.6 min | 4.6s | 57.0s |
| Mem0 |
1.9 min | 2.7s | 41.0s |
| Hyperspell |
3.1 min | 16.0s | 41.0s |
| gbrain |
3.2 min | local | 99.0s |
| Mitosis Cortex |
7.5 min | 0.40s | 55.0s |
| Zep |
9.1 min | 21.3s | 41.0s |
| Claude Code built-in memoryClaude Code memory | — | local | 25.0s |
Where retrieval time goesi
The average answer, second by second.
Let Your Agent Access This Data
Every number on this page is live over x402: an agent pays 0.035 USDC and gets the data with no account and no API key. A standard subscription based API is coming soon.
Pricing and cost
What the tools charge, and what successful answers actually cost.
What each tool costs to start, from list pricing.
| Tool | Free tier | Starting at | Open source |
|---|---|---|---|
| Karpathy Wiki |
Always | $0 | Open method |
| Mitosis Cortex |
7-day trial | $7.99/mo | Closed |
| gbrain |
Pay for embeddings | $0 | Yes |
| Hyperspell |
No cap hit at 200+ | Not Listed | Closed |
| Mem0 |
~550 answers/mo | $19/mo | Core is |
| Anthropic Memory |
Always | $0 | Yes |
| Supermemory |
~800 answers/mo | $19/mo | Yes |
| Zep |
~950 answers/mo | $125/mo | Engine (Graphiti) is |
| Claude Code built-in memoryClaude Code memory | Included with Claude Code | $0 | Bundled |
List prices as of Jul 27, 2026.
Cost Breakdowni
Per 1,000 successful answers: agent token spend, the entry paid tier amortized at its capacity, and embeddings where billed separately.