VRAM TableGPU fit and local-LLM VRAM for OEM desktops, from published specifications Search by model
Browse categories

GeForce RTX 2060 12GB (Reference design)

Reference design · 12 GB · catalogue identifier rtx-2060-12gb-ref · last updated 2026-09-24

82 desktops in this database carry a verdict for the GeForce RTX 2060 12GB (Reference design): 19 fit with conditions, 43 do not fit (all estimates) and 20 are unverified. Estimated Conditions most often concern card length (19 desktops), card width (19) and power (19); for this card, its length is not published and its slot width is not published. Of the 43 desktops it does not fit, 31 conflict with a half-height-only slot and 12 with the auxiliary power connector on record; the Dell OptiPlex 3090 Small Form Factor (2021) is one of the 31. Estimated Among the 17 cards with 12 GB in this database, its board power of 185 W is within 70–350 W, with 2 lower and 14 higher; its length is not published, while 14 cards in the group publish 168–285 mm. Estimated The largest listed model its 12 GB holds at 32k context is Gemma 3 12B (Q4_0 (QAT), needs 10.9 GiB). Estimated

Published specifications for this design, the OEM desktops whose own specifications accept it, and how much of a local language model its memory holds.

We have not physically tested this combination. Data is compiled from official specification sheets and cited sources.

“Reference design” is a specification, not a product. The figures below are the ones the chip vendor publishes for a baseline board: board power, auxiliary power connector and recommended system power. Cards sold by board partners are built to their own designs and differ in length, height and cooler. This database does not claim that a card built to this exact design was sold at retail, and where the chip vendor does not publish a dimension, the row below says so.

Specifications

Graphics processor
GeForce RTX 2060 12GB
Card memory
12 GB · D read 2026-09-19 [S1]
Total board power
185 W · E Source for this figure not recorded per field; see record sources
Vendor's recommended system power
450 W — a figure for the whole system, not for the card alone · E Source for this figure not recorded per field; see record sources
Auxiliary power connector
8pin · D read 2026-09-19 [S1]
Length
Not stated by the cited specification
Height
Not stated by the cited specification
Width
Not stated by the cited specification
Slot widths occupied
Not stated by the cited specification
Low profile
No · E Source for this figure not recorded per field; see record sources
CUDA
Yes

Each graded figure above cites: [S1] GeForce RTX 2060 12GB specifications (supporting source).

Power. Leave headroom; OEM PSUs are often rated at the limit. The recommended system figure above is the card vendor's own, and it assumes a retail supply rather than an OEM one.

Desktops that take a card of this design

This card has been assessed against 82 desktops in this database. Each verdict is a statement about published specifications, not about a tested combination.

  • ✓ Fits: 0
  • △ Fits with conditions: 19
  • ✕ Does not fit: 43 (all estimates) Estimated
  • ? Unverified: 20

Conditions shared by the desktops below. Each is stated once here. They come from this card's own records, or are worded the same for more than one desktop. A desktop listed below names the ones it carries by number; the figures particular to that desktop are on its own page.

  1. Card length: the length of this card is not published. Check it on the specification of the card you intend to buy. (19 desktops)
  2. Card length: where a desktop publishes no maximum, measure the clearance in front of the slot and compare it with the card. (11 desktops)
  3. Card width: the number of slot positions this card occupies is not published. Check it on the specification of the card you intend to buy. Card width: where a desktop publishes no slot-width ceiling, check that there is room beside the x16 slot. (19 desktops)
  4. Bracket height: where a row says whether a desktop takes only half-height cards is recorded from third-party or community material, or from material whose source the record does not identify, rather than from a statement by the manufacturer, check the slot opening. (2 desktops)
  5. Power: where a row says a desktop's supply rating is recorded from third-party or community material, or from material whose source the record does not identify, rather than from a specification the manufacturer publishes, read the rating off the label on the supply in the machine you own. (13 desktops)
  6. Power: this card's vendor recommends a 450 W system supply. Power: a card vendor's recommended system supply covers the whole machine and carries the vendor's own margin, so judge it against the configuration you own. (6 desktops)
  7. Auxiliary power connector: where a row says what a desktop's supply offers is recorded from third-party or community material, or from material whose source the record does not identify, rather than from a specification the manufacturer publishes, open the machine and look at the cable before buying. (8 desktops)
  8. Auxiliary power connector: this card needs an 8-pin connector. Auxiliary power connector: where a desktop's manufacturer does not publish what the supply offers, open the machine and look at the cable before buying. (11 desktops)

△ Fits with conditions 19 desktops:

✕ Does not fit 43 desktops, grouped by the published limit that conflicts with this card's published figures:

? Unverified 20 desktops. The evidence does not settle these either way, which is not the same as them being ruled out.

Verify with the manufacturer before purchasing.

What each verdict means. Every one of them is a statement about published specifications.
VerdictMeaning
✓ Fits Published specifications show no physical or electrical conflict.
△ Fits with conditions Published specifications allow it only under the stated conditions.
✕ Does not fit Published specifications rule the combination out: the card needs a connector the machine does not have, draws more than the supply provides, or is physically too large. It is a statement about specifications, not a report of a failed attempt.
? Unverified There is not enough evidence to say.

A dashed Estimated badge at the start of a line marks every verdict on that line as published as an estimate, and the grade beside it applies to each of them. The document behind each verdict, and the desktop's own figures, are listed on the desktop's own page. This card has a verdict for 82 of the 82 desktops in this database; the remaining 0 have not been assessed against it, which is not the same as them being ruled out.

Local language models this card's memory holds

Arithmetic shared by the models below. Sentences that read the same for more than one model are stated once here; each row names the ones it carries by number.

  1. This outcome is no stronger than the figures it was calculated from: this card's memory size is recorded here from third-party or community material, so grade D is shown alongside B. (5 rows)
  2. This outcome is no stronger than the figures it was calculated from: this card's memory size and the size of the model file that is distributed are recorded here from third-party or community material, so grade D is shown alongside B. (3 rows)
Calculated from the published file size and the model architecture.
ModelOutcome Needs at 32k contextThe arithmetic Confidence
Gemma 3 12B Q4_0 (QAT) ✓ Fits in VRAM Estimated 10.9 GiB 12 GB of memory is at least the 10.9 GB required at a context length of 32,768 tokens. [1] BD
Qwen3 8B Q4_K_M ✓ Fits in VRAM Estimated 10.2 GiB 12 GB of memory is at least the 10.2 GB required at a context length of 32,768 tokens. [1] BD
Llama 3.1 8B Q4_K_M ✓ Fits in VRAM Estimated 9.6 GiB 12 GB of memory is at least the 9.6 GB required at a context length of 32,768 tokens. This outcome is no stronger than the figures it was calculated from: this card's memory size and the size of the model file that is distributed and this model's layer count and this model's key/value head count and this model's head dimension are recorded here from third-party or community material, so grade D is shown alongside B. BD
Qwen3 4B Q4_K_M ✓ Fits in VRAM Estimated 7.9 GiB 12 GB of memory is at least the 7.9 GB required at a context length of 32,768 tokens. [1] BD
Qwen3 1.7B Q4_K_M ✓ Fits in VRAM Estimated 5.6 GiB 12 GB of memory is at least the 5.6 GB required at a context length of 32,768 tokens. [2] BD
Qwen3 0.6B Q4_K_M ✓ Fits in VRAM Estimated 4.9 GiB 12 GB of memory is at least the 4.9 GB required at a context length of 32,768 tokens. [2] BD
Gemma 3 4B Q4_0 (QAT) ✓ Fits in VRAM Estimated 4.7 GiB 12 GB of memory is at least the 4.7 GB required at a context length of 32,768 tokens. [1] BD
Phi-4 14B Q4_K_M △ Fits at short context only Estimated 15.6 GiB 12 GB of memory covers the 10.1 GB required at a context length of 4,096 tokens, but not the 15.6 GB required at 32,768 tokens. [2] BD
Qwen3 14B Q4_K_M △ Fits at short context only Estimated 14.4 GiB 12 GB of memory covers the 10.1 GB required at a context length of 4,096 tokens, but not the 14.4 GB required at 32,768 tokens. [1] BD

✕ Exceeds VRAM 9 models need more memory at 32k context than this card has, 9 of them published as an estimate. The arithmetic for each is on the model's own page.

Exceeds VRAM does not mean the model will not run: inference software can keep part of a model in system memory. How fast that is, and whether it is pleasant to use, is outside what this site measures. The sizing rule and its constants are on the Methodology page.

Cards whose published memory size is the same, so the local-model outcomes on this page apply to them as well. Whether a given desktop accepts them is a separate question, answered on each card's own page.

Sources

Every document this page draws on, with the date it was read.
DocumentPublisherKindRead on
GeForce RTX 2060 12GB specifications (supporting source) cputronic.com near-primary 2026-09-19
GeForce 20 series — Wikipedia en.wikipedia.org supporting 2026-09-19

We have not physically tested this combination. Data is compiled from official specification sheets and cited sources. Verify with the manufacturer before purchasing.