The hardware roster · one identity
the pair — two RTX 3080s in one 2021 desktop
whole machine·81 figures · 1 exhibit
two GeForce RTX 3080 10 GB cards of 2021 in a desktop of the same vintage, one at PCIe x16 and one at x4 of an x16-capable slot, on one supply — 20 GB of graphics memory between them, and the rig exhibit fifty measures against a single card in the same box
It holds — RTX 3080 (10 GB), Intel Core i9-10850K.
Narrow by model
Narrow by power cap
The rail is CSS and nothing else — no script runs on this page. Two things it therefore cannot do, said rather than discovered: a chosen filter is not part of the address, so a narrowed view cannot be linked or bookmarked; and it cannot tell you when nothing matches, which is why the count rides on every control and why a band whose rows are all filtered out keeps its heading and empties.
Language — decode, prefill and first token
Sorted by model, then by power cap ascending — alphabetical and numeric, never by value. Every condition the page states is printed; a condition it does not state reads not stated, which is a third thing and not an uncapped card. The columns a narrow screen cannot hold are printed under the row rather than scrolled to.
| value | what was measured | model | power cap | runtime | measurement window |
|---|---|---|---|---|---|
| 67.0 | decode per conversation, tokens a second | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out4two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 124.4 | decode summed across conversations, tokens a second | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out1two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 188.9 | decode summed across conversations, tokens a second | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out2two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 268.0 | decode summed across conversations, tokens a second | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out4two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 114.8 | decode, tokens a second | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M131,072256 tokens outnot statedthe same reading this page also prints in power-heat-and-the-fifty-wattstwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 2,531 | joules per 1,000 tokens | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 66.9 | decode per conversation, tokens a second | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out4two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 124.0 | decode summed across conversations, tokens a second | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out1two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 189.5 | decode summed across conversations, tokens a second | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out2two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 267.6 | decode summed across conversations, tokens a second | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens out4two-used-3080s-priced · the-full-concurrency-ladder · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 115.4 | decode, tokens a second | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M131,072256 tokens outnot statedthe same reading this page also prints in power-heat-and-the-fifty-wattstwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 2,994 | joules per 1,000 tokens | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 59.5 tok/s, 75.6 % on the cards | decode and residency out of the box, the server's own placement | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Ma 4,096 window256 tokens outnot statedtwo-used-3080s-priced · how-much-can-each-one-hold · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| out of memory — holds 65,536 instead, at 132.6 tok/s | decode at the full window, and the memory it used | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_M · f16, unquantisedits full 131,072-token window256 tokens outnot statedthe same reading this page also prints in how-much-can-each-one-holdtwo-used-3080s-priced · what-the-cache-type-costs · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 113.8 tok/s, 18,815 MiB | decode at the full window, and the memory it used | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_M · q4_0its full 131,072-token window256 tokens outnot statedtwo-used-3080s-priced · what-the-cache-type-costs · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 115.4 tok/s, 19,531 MiB | decode at the full window, and the memory it used | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_M · q8_0, the setting both rigs are measured atits full 131,072-token window256 tokens outnot statedthe same reading this page also prints in how-fast-does-each-one-writetwo-used-3080s-priced · what-the-cache-type-costs · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 124.7 | decode out of the box once the layer count is named | gemma4:26b | not stated | ollama · 0.32.13 | not stated |
| gemma4:26bnot statedollama · 0.32.13not stated | |||||
| Q4_K_Ma 4,096 window256 tokens outnot statedthe pair's own doubling, printed in the winner cell and nowhere else on the page; the same cell says the two rigs' out-of-the-box arms did not run at the same captwo-used-3080s-priced · at-a-glance · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 44.6 | decode, tokens a second | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M49,152256 tokens outnot statedthe same reading this page also prints in power-heat-and-the-fifty-wattstwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| does not fit | decode, tokens a second | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M65,536256 tokens outnot stated20 GB between two cards does not hold the dense model at this window; the pair's own ceiling is the 49,152 row abovetwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 10,442 | joules per 1,000 tokens | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 44.7 | decode, tokens a second | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M49,152256 tokens outnot statedthe same reading this page also prints in power-heat-and-the-fifty-wattstwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| does not fit | decode, tokens a second | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M65,536256 tokens outnot statedtwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 11,385 | joules per 1,000 tokens | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 61.5 tok/s | decode once the window is full, tokens a second | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_MOne 131,072-token window with about 98,000 tokens really in it256 tokens outnot statedtwo-used-3080s-priced · what-about-a-long-document · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 28.36 s | reading 98,024 tokens, seconds | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_MOne 131,072-token window with about 98,000 tokens really in it256 tokens outnot statedtwo-used-3080s-priced · what-about-a-long-document · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 30.04 s | time to the first token, seconds | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_MOne 131,072-token window with about 98,000 tokens really in it256 tokens outnot statedtwo-used-3080s-priced · what-about-a-long-document · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 94.8 | tokens a second each caller, two callers | one 26-billion model split across both cards | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| one 26-billion model split across both cards300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 189.5 | total tokens a second, two callers | one 26-billion model split across both cards | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| one 26-billion model split across both cards300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 59.1 | tokens a second each caller, two callers | two copies of a 12-billion model, one per card | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| two copies of a 12-billion model, one per card300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 118.2 | total tokens a second, two callers | two copies of a 12-billion model, one per card | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| two copies of a 12-billion model, one per card300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
The same card, the same question, concurrency apart
a gate-shaped call, median / p95
- 450 ms / 486 ms — 1 caller · 1 parallel slot two-used-3080s-priced · the-slot-trap-if-you-serve-other-people
- 709 ms / 712 ms — 1 caller · 4 parallel slots two-used-3080s-priced · the-slot-trap-if-you-serve-other-people
- 782 ms / 1,098 ms — 4 callers · 1 parallel slot two-used-3080s-priced · the-slot-trap-if-you-serve-other-people
- 1,076 ms / 7,599 ms — 4 callers · 4 parallel slots two-used-3080s-priced · the-slot-trap-if-you-serve-other-people
Every row above is the same card and the same mistral-small3.2:24b. The only thing that moved is the concurrency. That is why nothing on this page is ranked, and why a condition is part of a row rather than a heading above one.
The same card, the same question, power cap apart
core temp max
- 46 · 42 °C — 250 W each — 500 W between them two-used-3080s-priced · power-heat-and-the-fifty-watts
- 59 · 58 °C — 300 W each — 600 W between them two-used-3080s-priced · power-heat-and-the-fifty-watts
Every row above is the same card and the same gemma4:26b. The only thing that moved is the power cap. That is why nothing on this page is ranked, and why a condition is part of a row rather than a heading above one.
Pictures — seconds per render
2 rows — same rule as above.
| value | what was measured | model | power cap | runtime | measurement window |
|---|---|---|---|---|---|
| from 3.462 s to 2.261 s — 1.53× | one picture with the text encoder moved to the second card | not stated | 250 W each — 500 W between them | ComfyUI · 0.21.1 | 2026-09-17 |
| not stated250 W each — 500 W between themComfyUI · 0.21.12026-09-17 | |||||
| not statednot statednot statednot statedthe same bytes and the same picture — the page's own words, and it says this is more than the whole generational gap abovetwo-used-3080s-priced · renders-one-3080-against-one-3090-seconds-per-image · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 24.96 images a minute against 16.37 for one of them alone, 1.52× | two pictures at once, images a minute | not stated | 250 W each — 500 W between them | ComfyUI · 0.21.1 | 2026-09-17 |
| not stated250 W each — 500 W between themComfyUI · 0.21.12026-09-17 | |||||
| not statednot statednot statednot statedtwo-used-3080s-priced · renders-one-3080-against-one-3090-seconds-per-image · carried by hand, proved against that page’s own row · identity from the same page | |||||
Power, temperature, and what it costs
12 rows — same rule as above.
| value | what was measured | model | power cap | runtime | measurement window |
|---|---|---|---|---|---|
| 290 W | mean draw | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 40 to 46 °C | temperature on the quietest arm, with neither fan turning | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedthe page calls this the most surprising measurement on it: two five-year-old gaming cards writing in silencetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 346 W | mean draw | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 466 W | mean draw | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 0.2 per cent | what fifty watts are worth, dense | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot stated256 tokens outnot statedthe same reading this page also prints in power-heat-and-the-fifty-wattstwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 509 W | mean draw | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 0.5 per cent on the mixture and 0.2 on the dense model | what fifty watts buy on the pair | not stated | 250 W each — 500 W between them · 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated250 W each — 500 W between them · 300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedthe page's own heading for this reading is that on the pair, fifty watts buy nothingthe same reading this page also prints in how-fast-does-each-one-writetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 18 and 9 per cent more energy per thousand tokens | what fifty watts cost the pair in energy | not stated | 250 W each — 500 W between them · 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated250 W each — 500 W between them · 300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 5,837 | energy at depth, joules per 1,000 tokens | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_MOne 131,072-token window with about 98,000 tokens really in it256 tokens outnot statedtwo-used-3080s-priced · what-about-a-long-document · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| the pair spends 69.0 and 87.9 per cent more for the same thousand tokens | the exhibit's own energy cost against the single card | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 356 W | mean draw | one 26-billion model split across both cards | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| one 26-billion model split across both cards300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 542 W | mean draw | two copies of a 12-billion model, one per card | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| two copies of a 12-billion model, one per card300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
Memory — what fits, and at which context
1 row — same rule as above.
| value | what was measured | model | power cap | runtime | measurement window |
|---|---|---|---|---|---|
| 85.7 W | card board power, the model resident | not stated | 250 W each — 500 W between them | not stated | 2026-09-17, 01:43 to 01:50 |
| not stated250 W each — 500 W between themnot stated2026-09-17, 01:43 to 01:50 | |||||
| not statednot statednothing generatingnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
Everything else this shelf measured for it
36 rows — same rule as above.
| value | what was measured | model | power cap | runtime | measurement window |
|---|---|---|---|---|---|
| 46 · 42 °C | core temp max | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 0 % · 0 % | fan max | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 114.8 | tok/s | gemma4:26b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedthe same reading this page also prints in how-fast-does-each-one-writetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 59 · 58 °C | core temp max | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 64 % · 0 % | fan max | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 115.4 | tok/s | gemma4:26b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedthe same reading this page also prints in how-fast-does-each-one-writetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| holds 65,536 | largest window an unquantised cache leaves the pair | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_M · f16not statednot statednot statedthe same reading this page also prints in what-the-cache-type-coststwo-used-3080s-priced · how-much-can-each-one-hold · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 131,072 whole, 19,531 MiB used | largest window held whole, and the memory it used, the layer count named | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot stated256 tokens outnot statedtwo-used-3080s-priced · how-much-can-each-one-hold · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| about 5 GB of the pair's 20 unused | memory left unused under the server's own placement | gemma4:26b | not stated | ollama · 0.32.13 | 2026-09-17 |
| gemma4:26bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot stated256 tokens outnot statedtwo-used-3080s-priced · how-much-can-each-one-hold · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 58 · 52 °C | core temp max | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 62 % · 0 % | fan max | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 34.9 per cent | the exhibit's own margin over the single card, dense | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_M49,152256 tokens outnot statedthe page states this as the single card's loss; it is the pair's one decode win and the page calls it its turning pointtwo-used-3080s-priced · how-fast-does-each-one-write · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 44.6 | tok/s | mistral-small3.2:24b | 250 W each — 500 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b250 W each — 500 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedthe same reading this page also prints in how-fast-does-each-one-writetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 69 · 58 °C | core temp max | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 74 % · 0 % | fan max | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedper card, the re-padded Gigabyte board first and the stock EVGA board second — the page's own column notetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 44.7 | tok/s | mistral-small3.2:24b | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24b300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_Meach rig's headline window256 tokens outnot statedthe same reading this page also prints in how-fast-does-each-one-writetwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 450 ms / 486 ms | a gate-shaped call, median / p95 | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot statedeight tokens out1 caller · 1 parallel slottwo-used-3080s-priced · the-slot-trap-if-you-serve-other-people · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 709 ms / 712 ms | a gate-shaped call, median / p95 | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot statedeight tokens out1 caller · 4 parallel slotstwo-used-3080s-priced · the-slot-trap-if-you-serve-other-people · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 782 ms / 1,098 ms | a gate-shaped call, median / p95 | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot statedeight tokens out4 callers · 1 parallel slottwo-used-3080s-priced · the-slot-trap-if-you-serve-other-people · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 1,076 ms / 7,599 ms | a gate-shaped call, median / p95 | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot statedeight tokens out4 callers · 4 parallel slotstwo-used-3080s-priced · the-slot-trap-if-you-serve-other-people · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 49,152 whole, 18,909 MiB used | largest window held whole, and the memory it used, the layer count named | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot stated256 tokens outnot statedtwo-used-3080s-priced · how-much-can-each-one-hold · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 73.7 % | residency under the server's own placement | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot statedeight tokens out1 parallel slottwo-used-3080s-priced · the-slot-trap-if-you-serve-other-people · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| 41.7 % | residency under the server's own placement | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot statedeight tokens out4 parallel slotstwo-used-3080s-priced · the-slot-trap-if-you-serve-other-people · carried by hand, proved against that page’s own row · identity from the row itself | |||||
| never measured as a ladder; at a 49,152 window the planner kept 73.7 % on the cards | residency under the server's own placement | mistral-small3.2:24b | not stated | ollama · 0.32.13 | 2026-09-17 |
| mistral-small3.2:24bnot statedollama · 0.32.132026-09-17 | |||||
| Q4_K_Mnot stated256 tokens outnot statedtwo-used-3080s-priced · how-much-can-each-one-hold · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 44.0 W | card board power, cards empty | not stated | 250 W each — 500 W between them | not stated | 2026-09-17, 01:43 to 01:50 |
| not stated250 W each — 500 W between themnot stated2026-09-17, 01:43 to 01:50 | |||||
| not statednot statednothing generatingnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 0.78 kWh a day and 286 kWh a year | left on continuously, a day and a year | not stated | 250 W each — 500 W between them | not stated | 2026-09-17, 01:43 to 01:50 |
| not stated250 W each — 500 W between themnot stated2026-09-17, 01:43 to 01:50 | |||||
| not statednot statednothing generatingnot statedthe difference between the two rigs held idle, not either rig's own consumptiontwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| about $52 a year, or $4.37 a month | left on continuously, in money | not stated | 250 W each — 500 W between them | not stated | 2026-09-17, 01:43 to 01:50 |
| not stated250 W each — 500 W between themnot stated2026-09-17, 01:43 to 01:50 | |||||
| not statednot statednothing generatingnot statedat 18.34 cents a kilowatt-hour, the United States residential average for June 2026 as the page names it — and it is the DIFFERENCE between the two rigs held idle, card board power only, so a wall meter would read moretwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 41.7 W | what the model itself costs to hold | not stated | 250 W each — 500 W between them | not stated | 2026-09-17, 01:43 to 01:50 |
| not stated250 W each — 500 W between themnot stated2026-09-17, 01:43 to 01:50 | |||||
| not statednot statednothing generatingnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the table’s own column head | |||||
| 32.6 W more than the single card | what the pair costs over the single card to sit there holding a model | not stated | 250 W each — 500 W between them | not stated | 2026-09-17, 01:43 to 01:50 |
| not stated250 W each — 500 W between themnot stated2026-09-17, 01:43 to 01:50 | |||||
| not statednot statednothing generatingnot statedtwo-used-3080s-priced · power-heat-and-the-fifty-watts · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 1.59× | the exhibit's own ratio on the read | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | 2026-09-17 |
| not stated300 W each — 600 W between themollama · 0.32.132026-09-17 | |||||
| Q4_K_MOne 131,072-token window with about 98,000 tokens really in it256 tokens outnot statedtwo-used-3080s-priced · what-about-a-long-document · carried by hand, proved against that page’s own row · identity from the same page | |||||
| 60.3 per cent | the exhibit's own ratio, one big model against two copies | not stated | 300 W each — 600 W between them | ollama · 0.32.13 | Measured 2026-09-17 01:17 to 01:22 UTC |
| not stated300 W each — 600 W between themollama · 0.32.13Measured 2026-09-17 01:17 to 01:22 UTC | |||||
| Q4_K_Mnot stated256 tokens outtwo callerstwo-used-3080s-priced · two-copies-or-one-bigger-model · carried by hand, proved against that page’s own row · identity from the same page | |||||
| $1,483.80 | the build, total | not stated | not stated | not stated | read between 15:00 and 15:30 UTC on 17 September 2026 |
| not statednot statednot statedread between 15:00 and 15:30 UTC on 17 September 2026 | |||||
| not statednot statednot statednot statedthe record behind this figure is public — the listing its page prints in full, read in the window that page namestwo-used-3080s-priced · what-it-costs-priced-17-september-2026 · carried by hand, proved against that page’s own row · identity from the same page | |||||
| $1,488.81 | the build, total | not stated | not stated | not stated | read between 15:00 and 15:30 UTC on 17 September 2026 |
| not statednot statednot statedread between 15:00 and 15:30 UTC on 17 September 2026 | |||||
| not statednot statednot statednot statedthe record behind this figure is public — the listing its page prints in full, read in the window that page namestwo-used-3080s-priced · what-it-costs-priced-17-september-2026 · carried by hand, proved against that page’s own row · identity from the same page | |||||
| $1,578.80 | the build, total | not stated | not stated | not stated | read between 15:00 and 15:30 UTC on 17 September 2026 |
| not statednot statednot statedread between 15:00 and 15:30 UTC on 17 September 2026 | |||||
| not statednot statednot statednot statedthe record behind this figure is public — the listing its page prints in full, read in the window that page namestwo-used-3080s-priced · what-it-costs-priced-17-september-2026 · carried by hand, proved against that page’s own row · identity from the same page | |||||
| $1,080.83 | the build, total | not stated | not stated | not stated | read between 15:00 and 15:30 UTC on 17 September 2026 |
| not statednot statednot statedread between 15:00 and 15:30 UTC on 17 September 2026 | |||||
| not statednot statednot statednot statedthe record behind this figure is public — the listing its page prints in full, read in the window that page nameswhat a thousand dollars buys, and nothing else: no case, no drive, no operating system, and no shipping — the page's own list of what is still to addtwo-used-3080s-priced · what-it-costs-priced-17-september-2026 · carried by hand, proved against that page’s own row · identity from the same page | |||||
| $1,393.81 | the build, total | not stated | not stated | not stated | read between 15:00 and 15:30 UTC on 17 September 2026 |
| not statednot statednot statedread between 15:00 and 15:30 UTC on 17 September 2026 | |||||
| not statednot statednot statednot statedthe record behind this figure is public — the listing its page prints in full, read in the window that page namesthe cheapest combination the page totalstwo-used-3080s-priced · what-it-costs-priced-17-september-2026 · carried by hand, proved against that page’s own row · identity from the same page | |||||
Where it appears
Every released exhibit that names this identity, in the shelf’s own order. A count beside an exhibit is how many of its figures are carried above — and it links them, because a number with no way back to what it counts is the thing this shelf exists against.
- Two 3080s against one 3090, and the cap decides — 81 figures on this page
How this one is named here
The canonical name above is the full product name. Everything below is a spelling a published page or kit has actually used for it, kept so a reader who searches an older page’s words lands on this row. The set is closed and historical: no page cut after 2026-09-17 writes a memory class where it can write a card.
- two 3080s