Reliability Score · methodology v1.0 · updated 2026-10-04
Bitget: 74/100
Score over time
73 – 74 across 324 daily readings since 2026-06-23. The burgundy point is today.
Score breakdown
Withdrawal reliability weight 25%
100
Why 100
No recorded withdrawal halts or freezes within the 24-month decay window.
Proof of reserves weight 20%
80
Why 80 · 2 sources
Monthly self-published Merkle-tree PoR since 2023 with an open-source verification tool; Oct 2025 ratios BTC 307% / USDT 105%. No independent attestation documented.
Regulatory standing weight 20%
45
Why 45 · 2 sources
No MiCAR licence yet: FMA (Austria) application pending, EU HQ in Vienna with an EU CEO appointed Jan 2026; states it will not serve EEA customers until authorised. ASIC investor alert (unlicensed futures) and Japan FSA unregistered-exchange warnings are active.
Security history weight 15%
87
Why 87 · 1 source
1 recorded security incident within the 24-month window, weighted by severity, recency and reimbursement.
Support quality weight 10%
40
Why 40 · 1 source
Trustpilot 2.0/5 (~2,200 reviews, 2026-06-10) as a complaint-pattern proxy.
Operational maturity weight 10%
70
Why 70 · 1 source
Founded 2018; top-10 spot / top-tier derivatives venue with a $500M protection fund. The April 2025 VOXEL market-making-bot malfunction was contained with rollbacks within 24h (see incident timeline).
Incident timeline · facts with dates
2025-04-20
Security · sev 3
What happened · sourced
On 2025-04-20 a malfunction in Bitget’s own market-making bot on the VOXEL/USDT perpetual was exploited (~$100M extracted); irregular trades were rolled back within 24h with a compensation plan, and legal action was taken against eight accounts.
source
Sources & method: Method → · All exchanges: /exchanges
More on Bitget: hidden spread · withdrawal fees · move check: BTC→Bitcoin · ETH→Ethereum · ETH→Arbitrum One
Head to head: vs Binance · vs Bitpanda · vs Bitstamp · vs Bitvavo · vs Bybit
Not investment advice. Skontro is not a CASP under MiCAR. Reliability Scores are information, not recommendations: every negative datapoint is dated and sourced, the methodology is public, and no exchange can pay to change a score.