Gold & Metals

Silver fixes 2.26% higher as gold's fix slips 0.13%

LBMA published Friday's fixes for all four metals on 25 September. Platinum rose 0.87% and palladium fell 1.02%, while Kitco's screen closed higher on every one of the four.

By The Prices Desk

Corrected

What changed

London and the screen split. LBMA published Friday's fixes for all four metals. Silver's rose 2.26% and gold's fell 0.13% against Thursday. Kitco's screen closed higher on every one.

What it means

  • If you price off the fix: Silver's fix sits 1.32% above Friday's screen close of $64.169. Gold's sits 0.54% below the screen's $4,284.20.
  • If you settle against the screen: Kitco's own Friday session change reads gold up $10.80 and platinum up $23.00. Each is measured against Kitco's own prior close.
  • What this is not: A statement that the two benchmarks disagree about value. A fix is one auction print and a screen close is a bid at a stated minute, so a gap between them is expected.

Key figures

LBMA silver fix, 25 Sep+2.26% on 24 Sep65.015
LBMA gold PM fix, 25 Sep-0.13% on 24 Sep4,261.05
LBMA platinum PM fix+0.87% on 24 Sep1,768.80
LBMA palladium PM fix-1.02% on 24 Sep1,256.75

Source: LBMA fixes for 25 September 2026, read from the association's own JSON feeds on 26 September; screen levels from Kitco, stamped 2026-09-25T17:00:00Z

Two fixes moved against the screen

Gold and palladium fell. London's gold PM fix went from 4,266.40 to 4,261.05, and palladium's from 1,269.75 to 1,256.75.

The screen went up. Kitco's board shows Friday closes above its own prior closes on all four metals, gold by $10.80.

Friday's screen close against Friday's London fixKitco bid, stamped 2026-09-25T17:00:00Z; LBMA fixes dated 2026-09-25
MetalScreen close, 25 Sep (USD)LBMA fix, 25 Sep (USD)Fix minus screenFix against screen
Gold4,284.204,261.05-23.15-0.54%
Silver64.16965.015+0.846+1.32%
Platinum1,773.001,768.80-4.20-0.24%
Palladium1,251.001,256.75+5.75+0.46%
Notes on this table

What is being compared two different instruments on the same date. The screen column is Kitco's bid, the level the tape carries for Friday, stamped by Kitco's own feed at 2026-09-25T17:00:00Z. The fix column is LBMA's published price for 25 September, the PM auction for gold, platinum and palladium and the single daily auction for silver.

The arithmetic, shown both ways gold's fix of 4,261.05 divided by the screen's 4,284.20 is 99.46% of it, therefore 0.54% below it. Silver's 65.015 divided by 64.169 is 101.32%, therefore 1.32% above. Platinum's 1,768.80 divided by 1,773.00 is 99.76%, therefore 0.24% below. Palladium's 1,256.75 divided by 1,251.00 is 100.46%, therefore 0.46% above.

The day-on-day figures are computed on LBMA's own consecutive points, 24 to 25 September: gold 4,266.40 to 4,261.05, silver 63.58 to 65.015, platinum 1,753.60 to 1,768.80, palladium 1,269.75 to 1,256.75.

Precision. LBMA publishes gold, platinum and palladium to two decimals and silver to three, and no figure here is carried beyond the source's own precision. The percentages are the desk's division of those published levels.

Kitco's own change field gold +10.80, silver +0.445, platinum +23.00, palladium +4.00, each measured by Kitco against Kitco's own prior close and printed here as Kitco's figure, never used to derive a prior mark.

Silver made the largest move

A 2.26% fix gain. Silver's fix rose from 63.58 on 24 September to 65.015 on 25 September, the largest of the four moves.

The screen read lower. The tape carries silver's Friday close at $64.169, which is 1.32% under London's fix.

The board has not ticked

The board is frozen. Kitco stamps every metal 2026-09-25T17:00:00Z, identical across all four, so no new session has printed.

One price stands. Those four levels are the marks the tape carries for Friday, and they are the levels printed here.

What to watch

  • Monday 28 SeptemberLBMA's next set of fixes, published as dated JSON at prices.lbma.org.uk.
  • Next screen printKitco's timestamp moving off 2026-09-25T17:00:00Z, which dates the next session rather than any calendar.
  • Early OctoberRapaport's next RAPI points, the tape's other dated benchmark.

The story so far

  1. Gold rises $42.10 to $4,296.60 after Thursday's four-metal fall
  2. Silver adds 4.87% to $66.835, highest mark since 31 August

Go deeper

Corrections1 note

Corrected 27 September 2026. As published on 26 September this article used the phrase "the level this paper’s tape carries for Friday" in its table note, a Layer 1-2 surface. The article doctrine bans self-reference of that kind outside the depth, and the 26 September edition log asserted that a check for those phrases across every Layer 1-2 surface had returned zero hits. That assertion was false: three of the five articles in that edition carried one such phrase each, this being one of them. The phrase now reads "the level the tape carries for Friday". No figure in this article was wrong as published and none has changed: every number printed here was correct on 26 September, traces to the same cited source, and stands. The standards desk found the fault on 26 September 2026 and it is corrected here the next morning; the edition log's false zero-hits claim is corrected in the same run.

What would change this call

A revision. LBMA republishes its series as dated JSON and a point can be restated; the levels here were read on 26 September and a later read could differ. A Kitco timestamp that moves without a level moving, or a level that moves without the timestamp moving, would also change the reading, because both were used to establish that Friday's close is final rather than intraday.

A fix and a screen measure different things

The London fix is the clearing price of an electronic auction run at a set time, with a defined set of participants entering buy and sell volumes until the imbalance falls inside a tolerance. It is one number for one moment, and it exists because a contract needs a single reference both sides accept. A screen close is not an auction at all: it is the last bid a quote vendor carries before the market pauses, sampled from over-the-counter dealing that has no closing bell. The two therefore answer different questions. The fix answers what price cleared a book at a stated minute in London. The screen answers what a dealer was bidding at a stated minute somewhere. A gap between them on the same date is normal and its size is information about the hours between the two prints, not evidence that either is wrong. It becomes worth reading when the gap changes sign across metals in one session, as it does here, with two metals fixing above the screen and two below.

Method

All four LBMA feeds were fetched directly as raw JSON rather than through any summary, because a summarising read has produced invented digits on this desk before: prices.lbma.org.uk/json/gold_pm.json, silver.json, platinum_pm.json and palladium_pm.json, four separate 200 responses on 26 September 2026. Each was parsed and its newest points printed, and each carries a 25 September 2026 entry, so the claim that LBMA published Friday's fixes for all four metals rests on all four feeds and not on one of them. Their newest values are gold PM 4,261.05, silver 65.015, platinum PM 1,768.80 and palladium PM 1,256.75, with the 24 September points at 4,266.40, 63.58, 1,753.60 and 1,269.75. Kitco's board was re-read at press time on 26 September under the one-price rule, and the four bids it returned are identical to the marks the tape already carried for Friday: gold 4,284.20, silver 64.169, platinum 1,773.00, palladium 1,251.00. There is no divergence between the press-time read and the tape, so nothing was written back and the tape's own marks print here unchanged. Kitco's per-metal originalTime field reads 2026-09-25T17:00:00Z on all four, which is the evidence for saying the board has not ticked; no closure or holiday is asserted anywhere in this article, and the frozen-timestamp claim is read off the feed rather than off a calendar. The percentages are the desk's own division of the published levels and are shown in the table note so the base is visible in each case. Against this paper's 25 September metals article, which reported a press-time screen reading taken on Friday morning, the period here is different: that piece covered Friday's opening marks and this one covers Friday's settled close and Friday's published London fixes, which did not exist when it ran. Same instruments, a different window, so this advances rather than repeats.

Sources5