Every day, thousands of tokens are launched on Pump.fun. Roughly 1-2% of them ever complete the bonding curve and graduate to PumpSwap. The rest die on the curve, taking buyer SOL with them. Much of that fate is decided fast — we break down the timing in the first 3 minutes: how 1-in-200 pump.fun tokens graduate.
That 1-2% stat hides something important: graduation is not uniformly distributed across deployers. A small number of wallets produce graduates at rates 30x to 70x the baseline. If you can identify those wallets before they launch their next token, you have a measurable edge over every trader who is still picking tokens by name, narrative, or Telegram hype.
Deployer Hunter tracks 10,600+ pump.fun deployer wallets, classifies them into tiers based on historical performance, and fires real-time alerts when the best ones launch something new. This article explains exactly how the scoring works, what the numbers look like, and how traders can use the system.
The Data Pipeline: From On-Chain Event to Deployer Score
Before we get into the scoring logic, it helps to understand where the data comes from. Every deployer stat is derived from raw on-chain events, not scraped from third-party APIs.
Token creation detection. A Yellowstone gRPC subscription monitors the pump.fun program in real-time. Every time a wallet creates a new token, the system captures the deployer address, token mint, and block timestamp. The deployer's profile is created (if new) or updated within seconds.
Bond detection. When a token accumulates approximately 85 SOL in the bonding curve (the exact USD market cap at graduation floats with the SOL price), pump.fun migrates it to PumpSwap — their own AMM. The system detects this by subscribing to PumpSwap pool creation events via gRPC. Once a pool creation is confirmed, the deployer's bond count increments and their stats recalculate.
Market cap at bond time. An internal mc-tracker service provides live market cap data at the moment of bonding. This is how we know what price a token graduated at — not from a third-party API call minutes later, but from our own price feed at the exact block.
Daily snapshots. The system takes daily snapshots of every deployer's stats. This creates a trajectory: you can see whether a deployer is getting better or worse over weeks and months. A deployer whose bond rate has climbed from 20% to 45% over 30 days is a different signal than one whose rate has been flat at 45% for six months.
All of this feeds into the tier classification engine.
The Tier System: How 10,600 Deployers Get Classified
Every deployer in the system is classified into one of four tiers: Elite, Good, Rising, or New. The classification runs after every deploy and bond event — tiers are not static labels but live calculations.
Here is how each tier works.
Elite: The Top 34
To qualify as Elite, a deployer needs:
- 5 or more total launches (sufficient sample size)
- 40% or higher lifetime bonding rate (at least 2 in 5 tokens graduate)
- Strong recent performance (last 10 outcomes weighted)
- Active streaks (consecutive bonds boost confidence)
As of today, exactly 34 deployers out of 10,600+ meet these criteria. That is 0.3% of the tracked population. Their aggregate bonding rate is 71% — meaning roughly 7 out of every 10 tokens they launch complete the bonding curve and migrate to PumpSwap.
Compare that to the platform-wide baseline of 1-2%. Elite deployers graduate tokens at roughly 35x to 70x the average rate.
These are not lucky one-hit wallets. A deployer cannot reach Elite with a small sample. Five launches is the minimum, and at a 40% floor, that means at least two successful graduates with strong recent outcomes to back it up. Most Elite deployers have 10+ launches and bond rates between 50% and 90%.
Good: Consistent Performers
Good-tier deployers have:
- 3 or more launches
- 25% or higher lifetime bonding rate
- Recent outcomes that confirm the pattern (not just historical success from months ago)
This tier captures deployers who are clearly above average but do not yet have the sustained track record or volume to qualify as Elite. Many Good deployers are on their way up — the trajectory snapshots show a rising curve. Some will reach Elite within a few weeks if their next launches bond.
Rising: Small Sample, Perfect Record
Rising deployers have launched between 1 and 3 tokens, and every single one bonded. A 100% rate on a small sample.
This tier exists because a deployer with 2 launches and 2 bonds has a genuinely strong signal — it is just statistically fragile. The system surfaces them but with the implicit caveat that more data is needed. If their third token dies, they drop. If it bonds, they graduate to Good.
New: Insufficient Data
Every deployer who does not meet the criteria for Elite, Good, or Rising lands here. This includes:
- First-time deployers with no history
- Deployers with multiple launches but low or zero bond rates
- Previously ranked deployers who have gone cold (last 10 outcomes all dead)
New-tier deployers do not trigger alerts. They exist in the system for completeness, and they can climb into a tracked tier if their outcomes improve.
The Scoring Factors in Detail
The tier labels are the visible output. Behind them are four specific factors the system evaluates.
Factor 1: Total Launch Count
Raw volume matters for statistical confidence. A 50% bond rate on 2 launches is noise. A 50% bond rate on 20 launches is a pattern.
The thresholds are deliberately conservative:
- Elite requires 5+ launches
- Good requires 3+ launches
- Rising accepts 1-3 launches but only at 100%
These minimums filter out lucky wallets. Anyone can get one token to bond. Doing it consistently across 5, 10, or 20 launches requires something more — whether that is community-building skill, marketing execution, or deep understanding of what narratives resonate.
Factor 2: Lifetime Bonding Rate
This is the headline number: total bonds divided by total launches, expressed as a percentage.
The platform baseline is 1-2%. Here is how the tiers compare:
| Tier | Deployers | Avg Bond Rate | vs. Baseline |
|---|
| Elite | 34 | 71% | ~35-70x |
| Good | ~150-200 | 30-45% | ~15-40x |
| Rising | ~50-100 | 100% (small n) | ~50-100x |
| New | 10,000+ | varies widely | — |
The separation between tiers is large. An Elite deployer at 71% is not marginally better than baseline — it is operating in a fundamentally different regime.
Factor 3: Recent Outcomes (Last 10)
Lifetime rate can be misleading. A deployer who bonded 8 of their first 10 tokens but then launched 15 straight failures still has a 32% lifetime rate (8/25). Technically above baseline, but the current signal is dead.
To handle this, the system tracks the last 10 outcomes as an ordered array. Each entry is either B (bonded) or D (died). For example:
B B B D B B B D B B — strong recent performance, 8 of last 10 bondedD D D D D D D D D D — cold streak, last 10 all deadB D B D B D B D B D — inconsistent, 50% recent rate
The recent outcomes array is weighted in the tier calculation. A deployer whose lifetime rate qualifies for Good but whose last 10 are all D will drop. Conversely, a deployer whose lifetime rate is borderline but whose last 10 show 7 or 8 bonds gets a boost.
This recency weighting ensures the tiers reflect current deployer performance, not historical glory. For a step-by-step trader workflow built on these recent-outcome signals, see our guide on how to find winning tokens before they bond.
Factor 4: Streaks
Consecutive bonds are a stronger signal than the same bond rate with alternating outcomes. A deployer on a 5-bond streak is different from one with 5 bonds scattered across 10 launches.
The system tracks both current streak (consecutive bonds from the most recent launch backward) and best streak (longest consecutive bond run ever). Streaks feed into the tier classification as a confidence modifier: at the boundary between tiers, a strong streak can tip a deployer up.
Trajectory Tracking: How Deployers Change Over Time
Static snapshots miss an important dimension: direction. A deployer with a 35% bond rate today could be:
- Improving — started at 15%, now at 35% and climbing
- Stable — has been at 33-37% for three months
- Declining — peaked at 55%, now sliding toward 35%
The system captures daily snapshots of every deployer's stats — bond rate, total launches, tier, recent outcomes. These snapshots create a trajectory line that shows skill progression (or regression) over time.
For traders, trajectory is actionable intelligence. A Good-tier deployer on an upward trajectory is a better bet than a Good-tier deployer on a downward one, even if their current numbers are identical. The daily snapshots make this visible.
Real-Time Alerts: gRPC-Powered, Sub-Second
When an Elite, Good, or Rising deployer launches a new token, the alert fires within seconds. Here is the pipeline:
- gRPC subscription detects the pump.fun
create instruction on-chain
- Deployer lookup checks the wallet against the 10,600+ deployer database
- Tier check — if the deployer is Elite, Good, or Rising, an alert is generated
- Alert enrichment — the alert includes deployer tier, lifetime stats, recent outcomes, current streak, and the new token's mint address
- Distribution — the alert appears on the Deployer Hunter live feed and is available via the paid API
The same pipeline fires for bond events. When a tracked deployer's token graduates to PumpSwap, a bond alert is generated with the market cap at bond time (from mc-tracker), time to bond, and the deployer's updated stats.
Total latency from on-chain event to alert: typically under 2 seconds. There is no polling, no scraping, and no third-party API in the critical path.
How Traders Should Use This Data
The tier system is a filter, not a buy signal. Here is a practical framework for using it.
Prioritize Elite and Good Tiers
When an Elite deployer launches a token, 7 out of 10 times it will bond. That does not mean you should blindly buy every Elite launch. It means the token has cleared the most important statistical hurdle: the deployer behind it has a demonstrated history of producing tokens that graduate. For the practical, trader-facing workflow built on this scoring, see our guide on using the Pump.fun deployer tracker to find winning tokens before they bond.
