Lowest Spread Forex Brokers 2026: 19 Accounts Tested

Spread Statistics for All 19 Forex Brokers

Explore key spread statistics for each Forex broker in our test.

How to Choose a Low-Spread Forex Broker

A useful spread comparison starts with how you actually trade. These three checks help turn the study results into a practical shortlist.

Start with the pairs you trade

If most of your volume is in EURUSD, its column is more relevant to you than the 7-pair avg. Build your shortlist around your own pair mix.

Compare the same pricing model

These are Standard and standard-equivalent accounts. A Raw or Zero account may quote a tighter spread but charge commission, so add that commission to its cost before you compare. Confirm the exact account, platform, and legal entity available where you live.

Calculate the cost of your trade

Multiply the spread in pips by the pip value for your position size, then add any commission. For positions held longer, include swaps or financing and currency conversion; for active trading, allow for slippage as well.

Forex Spread Analysis for All 19 Brokers

Using the 7-pair avg. and the other spread statistics, here's where each broker's spreads were tighter, where they were wider, and how consistent they stayed across the seven major pairs during the test.

How We Tested and Ranked Forex Broker Spreads

The original collection files remain unchanged, and the public dataset lets you inspect the synchronized observations, account scope, and validation checks behind the tables.

How We Collected and Validated the Data

1. Recorded one quote per minute. Data collection ran from 9 July 2026, 10:00:30–19:00:30 UTC. In the collection timezone, that was 9 July 2026, 17:00:30–10 July 2026, 02:00:30 ICT (UTC+7). Spread values were normalized to pips before aggregation. The 18 MetaTrader 5 setups used collection version 1.1.6-mt5. XTB used xStation 5 Web with version 1.0.0. Both outputs were normalized to the same UTC timestamp, symbol, Bid, Ask, and spread fields before comparison.
2. Kept missing readings missing. Only rows with status=ok were retained. We did not impute, interpolate, carry forward, or replace missing spread values. The analysis rejected duplicate keys. The ranking script did not rewrite the raw files.
3. Matched the same minutes. For each symbol, a timestamp was eligible only when every broker in the original synchronized collection had a valid spread at that timestamp. We froze that timestamp set before ranking, so no published account benefited from extra minutes. Depending on the pair, 506 to 510 common timestamps remained. Average retained coverage across the seven primary pairs was 94.0% of the planned timestamps; every pair still received one-seventh of the overall score.

Publication checks for this study: 19 published brokers, 19 cleaned broker files, seven ranked major pairs, and 67,659 retained major-pair observations. Every displayed account had a verified official source and test setup, source hashes remained unchanged, and the synchronized observations passed validation. If a required check fails, the study is not published.

Tested Accounts, Platforms, and Servers

Each record below shows the account type, platform, and server used for that broker. The screenshots document the tested setup; the spread statistics come from the synchronized CSV dataset, not from a single image.

Limits of Our Lowest-Spread Forex Broker Test

1. The results are a snapshot, not a permanent league table. Spreads respond to market liquidity and volatility, and a single July trading window cannot represent every session, rollover, news event, holiday, or stressed market. Repeating the test over more days and sessions would show whether the rank differences hold.
2. Platform and entity context also matter. A broker can operate through multiple legal entities and offer different account names, symbols, markups, or product terms by jurisdiction. Do not assume a quote from the tested setup applies on every platform or in every country. Collecting data from an account does not show that it is available to you; check your own eligibility and the entity that would hold your account.
3. We used live trading accounts for all 19 brokers, although the collection platforms differed. 18 accounts were observed in MetaTrader 5, while XTB was observed in xStation 5 Web. Matching the data fields and timestamps reduces one source of mismatch, but it cannot remove platform-specific differences in how quotes are displayed or delivered.
4. We measured quoted bid/ask spread, not completed trades. The test does not measure fill price, rejected orders, requotes, latency, positive or negative slippage, execution speed, depth, or the difference between a displayed quote and an executable price at a chosen trade size. Any of these can change what you actually pay, even when the quoted spread looks similar.
5. Spread is not the same as total trading cost. Commission, swaps or overnight financing, currency conversion, inactivity charges, deposits and withdrawals, taxes, and instrument-specific fees are outside the dataset, so we do not call any result the cheapest broker overall. If you hold positions overnight or use a commission-based account, you need a fuller cost model than this spread-only comparison.
6. Small differences need context. Precise measurement does not make every gap meaningful. A gap of a few hundredths of a pip in this window may reverse in another session. Read the 7-pair avg., Median, P95, IQR, pair-level results, and study limitations together instead of treating rank alone as a prediction.

Download the 2026 Forex Broker Spread Dataset

Download all 67,659 synchronized observations used in this ranking, including timestamp, broker, currency pair, Bid, Ask, and spread in pips. The ZIP also contains one CSV per broker, aggregated results, methodology, a sanitized cleaning summary, and SHA-256 checksums.

Account numbers, account companies, servers, logins, local paths, and internal logger fields are excluded. Study version: 2026-07-09-19-brokers-standard-accounts.

Lowest Spread Forex Brokers: FAQs