llmrun

Gemma 4 26B A4B IT vs GPT OSS 20B

Gemma 4 26B A4B IT leads llmrun Score 105.4 to 96.2; both fit 24 GB at Q4_K_M.

Model A · Google

Gemma 4 26B A4B IT
105.4llmrun Score

25.8B · 4.4B active · 256K context · Apache 2.0 · Mar 2026

VisionFunctions

Model B · OpenAI

GPT OSS 20B
96.2llmrun Score

20.9B · 4.2B active · 128K context · Apache 2.0 · Aug 2025

ChatFunctions

Which should you choose?

Choose Gemma 4 26B A4B IT if you want…

  • Higher llmrun Score

    105.4 vs 96.2

  • Stronger at reasoning

    100.8 vs 90.8

  • Stronger at science

    102.2 vs 92.4

  • Ahead on the benchmarks

    7 of 9 shared boards

  • Longer context window

    256K vs 128K tokens

  • Also does vision

    GPT OSS 20B doesn't list it

Choose GPT OSS 20B if you want…

  • Lighter on memory

    12.9 vs 16.1 GB at Q4_K_M

  • Also does chat

    Gemma 4 26B A4B IT doesn't list it

On your hardware

Each model at its default quant (Q4_K_M), a smaller one when it must. Estimates, ±20%.

8 GB

RTX 4060 · 8 GB tier

Gemma 4 26B A4B IT

Doesn't fit

smallest quant needs 8.1 GB

GPT OSS 20B

IQ2_M · ~103 tok/s

7.6 GB · up to 32Kfits

12 GB

RTX 3060 12GB · 12 GB tier

Gemma 4 26B A4B IT

IQ3_S · ~99 tok/s

12 GB · up to 16K

GPT OSS 20B

IQ4_XS · ~89 tok/s

11.8 GB · up to 16Kbetter quant

16 GB

RTX 5060 Ti 16GB · 16 GB tier

Gemma 4 26B A4B IT

Q4_1 · ~95 tok/s

15.5 GB · up to 16K

GPT OSS 20B

Q4_K_M · ~98 tok/s

13.1 GB · up to 128Kmore headroom

24 GBreference

RTX 4090 · 24 GB tier

Gemma 4 26B A4B IT

Q4_K_M · ~171 tok/s

16.5 GB · up to 128K

GPT OSS 20B

Q4_K_M · ~190 tok/s

13.1 GB · up to 128Kfaster

32 GB

RTX 5090 · 32 GB tier

Gemma 4 26B A4B IT

Q4_K_M · ~250 tok/s

16.5 GB · up to 256K

GPT OSS 20B

Q4_K_M · ~282 tok/s

13.1 GB · up to 128Kfaster

48 GB

RTX A6000 · 48 GB tier

Gemma 4 26B A4B IT

Q4_K_M · ~140 tok/s

16.5 GB · up to 256K

GPT OSS 20B

Q4_K_M · ~154 tok/s

13.1 GB · up to 128Kfaster

Mac 32 GB

Mac Mini M4 · Macs

Gemma 4 26B A4B IT

Q4_K_M · ~26 tok/s

16.5 GB · up to 256K

GPT OSS 20B

Q4_K_M · ~28 tok/s

13.1 GB · up to 128K

Mac 64 GB

Mac Studio M4 Max · Macs

Gemma 4 26B A4B IT

Q4_K_M · ~76 tok/s

16.5 GB · up to 256K

GPT OSS 20B

Q4_K_M · ~86 tok/s

13.1 GB · up to 128Kfaster

Mac 128 GB

Mac Studio M4 Max · Macs

Gemma 4 26B A4B IT

Q4_K_M · ~76 tok/s

16.5 GB · up to 256K

GPT OSS 20B

Q4_K_M · ~86 tok/s

13.1 GB · up to 128Kfaster

Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).

llmrun Score

with its 90% range · bars share one scale

Overall

Gemma 4 26B A4B IT

105.4102–109

GPT OSS 20B

96.293–99

Coding

Gemma 4 26B A4B IT

103.695–112

GPT OSS 20B

100.992–110

Science

Gemma 4 26B A4B IT

102.295–110

GPT OSS 20B

92.485–100

Reasoning

Gemma 4 26B A4B IT

100.893–109

GPT OSS 20B

90.882–99

Gemma 4 26B A4B IT · #37 of 92 open models · #6 of 34 that fit 24 GB

GPT OSS 20B · #58 of 92 open models · #15 of 34 that fit 24 GB

Benchmarks

9 shared boards: Gemma 4 26B A4B IT leads 7, GPT OSS 20B leads 2 · Benchmarks imported Oct 2026

GPQA Diamond

Gemma 4 26B A4B IT

73.2%#29/97 open

IndependentEpoch AI · Apr 2026

GPT OSS 20B

60.8%#41/97 open

IndependentEpoch AI · Aug 2025

LMArena Text

Gemma 4 26B A4B IT

1437.6#23/186 open

ArenaLMArena · Oct 2026

GPT OSS 20B

1317.3#96/186 open

ArenaLMArena · Oct 2026

AIME 2024/2025

Gemma 4 26B A4B IT

82.2%#23/85 open

IndependentEpoch AI · Apr 2026

GPT OSS 20B

65.3%#33/85 open

IndependentEpoch AI · Aug 2025

SciCode

Gemma 4 26B A4B IT

40.1%#30/65 open

Lab-reportedEpoch AI · Apr 2026

GPT OSS 20B

34.4%#47/65 open

Lab-reportedEpoch AI · Aug 2025

WeirdML

Gemma 4 26B A4B IT

35.2%#34/46 open

IndependentEpoch AI · Apr 2026

GPT OSS 20B

40.9%#21/46 open

IndependentEpoch AI · Aug 2025

Chess Puzzles

Gemma 4 26B A4B IT

6.0%#27/67 open

IndependentEpoch AI · Apr 2026

GPT OSS 20B

4.0%#33/67 open

IndependentEpoch AI · Aug 2025

CritPt

Gemma 4 26B A4B IT

0.0%#50/65 open

Lab-reportedEpoch AI · Apr 2026

GPT OSS 20B

1.4%#30/65 open

Lab-reportedEpoch AI · Aug 2025

DTBench

Gemma 4 26B A4B IT

74.9%#29/68 open

Lab-reportedEpoch AI · Apr 2026

GPT OSS 20B

68.0%#34/68 open

Lab-reportedEpoch AI · Aug 2025

LMCA

Gemma 4 26B A4B IT

29.7%#20/54 open

Lab-reportedEpoch AI · Apr 2026

GPT OSS 20B

14.5%#43/54 open

Lab-reportedEpoch AI · Aug 2025

4 more boards (other shared and single-model boards)
ALE-Bench

Gemma 4 26B A4B IT

927.2#11/42 open

Lab-reportedEpoch AI · Apr 2026

GPT OSS 20B

566.0#29/42 open

Lab-reportedEpoch AI · Aug 2025

Only Gemma 4 26B A4B IT (1)

Only GPT OSS 20B (2)

VRAM by quantization

Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.

Q2_K

Gemma 4 26B A4B IT

11.6 GB11.0 GB file

32K 12.5 GB · 128K 15.8 GB

GPT OSS 20B

9.3 GB11.5 GB file

32K 9.8 GB · 128K 11.5 GB

Q3_K_S

Gemma 4 26B A4B IT

11.9 GB12.6 GB file

32K 12.8 GB · 128K 16.2 GB

GPT OSS 20B

9.5 GB11.5 GB file

32K 10.0 GB · 128K 11.7 GB

Q3_K_M

Gemma 4 26B A4B IT

13.2 GB12.7 GB file

32K 14.1 GB · 128K 17.5 GB

GPT OSS 20B

10.6 GB11.5 GB file

32K 11.1 GB · 128K 12.8 GB

Q4_0

Gemma 4 26B A4B IT

13.6 GB~12.9 GB file

32K 14.5 GB · 128K 17.8 GB

GPT OSS 20B

10.8 GB11.5 GB file

32K 11.3 GB · 128K 13.0 GB

Q4_K_M

Gemma 4 26B A4B IT

16.1 GB16.9 GB file

32K 17.0 GB · 128K 20.4 GB

GPT OSS 20B

12.9 GB11.6 GB file

32K 13.4 GB · 128K 15.1 GB

Q5_K_M

Gemma 4 26B A4B IT

19.0 GB21.2 GB file

32K 19.9 GB · 128K 23.3 GB

GPT OSS 20B

15.3 GB11.7 GB file

32K 15.8 GB · 128K 17.5 GB

Q6_K

Gemma 4 26B A4B IT

21.9 GB23.2 GB file

32K 22.8 GB · 128K 26.2 GB

GPT OSS 20B

17.6 GB12.0 GB file

32K 18.1 GB · 128K 19.8 GB

Q8_0

Gemma 4 26B A4B IT

26.5 GB~25.8 GB file

32K 27.4 GB · 128K 30.7 GB

GPT OSS 20B

21.3 GB~20.9 GB file

32K 21.8 GB · 128K 23.5 GB

19 more quants

IQ2_XXS

Gemma 4 26B A4B IT

7.7 GB9.9 GB file

32K 8.6 GB · 128K 12.0 GB

GPT OSS 20B

6.1 GB~5.8 GB file

32K 6.6 GB · 128K 8.3 GB

IQ2_XS

Gemma 4 26B A4B IT

8.4 GB10.1 GB file

32K 9.3 GB · 128K 12.6 GB

GPT OSS 20B

6.7 GB~6.3 GB file

32K 7.2 GB · 128K 8.9 GB

IQ2_S

Gemma 4 26B A4B IT

8.7 GB10.2 GB file

32K 9.6 GB · 128K 12.9 GB

GPT OSS 20B

6.9 GB~6.5 GB file

32K 7.4 GB · 128K 9.1 GB

IQ2_M

Gemma 4 26B A4B IT

9.4 GB10.0 GB file

32K 10.3 GB · 128K 13.6 GB

GPT OSS 20B

7.4 GB~7.1 GB file

32K 7.9 GB · 128K 9.6 GB

IQ3_XXS

Gemma 4 26B A4B IT

10.7 GB11.4 GB file

32K 11.6 GB · 128K 14.9 GB

GPT OSS 20B

8.5 GB11.6 GB file

32K 9.0 GB · 128K 10.7 GB

Q2_K_S

Gemma 4 26B A4B IT

11.0 GB10.6 GB file

32K 11.9 GB · 128K 15.2 GB

GPT OSS 20B

8.7 GB12.1 GB file

32K 9.2 GB · 128K 10.9 GB

IQ3_XS

Gemma 4 26B A4B IT

11.3 GB12.4 GB file

32K 12.2 GB · 128K 15.5 GB

GPT OSS 20B

9.0 GB11.6 GB file

32K 9.5 GB · 128K 11.2 GB

IQ3_S

Gemma 4 26B A4B IT

11.6 GB11.3 GB file

32K 12.5 GB · 128K 15.8 GB

GPT OSS 20B

9.3 GB12.1 GB file

32K 9.8 GB · 128K 11.5 GB

IQ3_M

Gemma 4 26B A4B IT

12.3 GB13.3 GB file

32K 13.2 GB · 128K 16.5 GB

GPT OSS 20B

9.8 GB11.6 GB file

32K 10.3 GB · 128K 12.0 GB

Q3_K_L

Gemma 4 26B A4B IT

13.9 GB13.2 GB file

32K 14.8 GB · 128K 18.1 GB

GPT OSS 20B

11.1 GB11.5 GB file

32K 11.6 GB · 128K 13.3 GB

IQ4_XS

Gemma 4 26B A4B IT

14.5 GB13.6 GB file

32K 15.4 GB · 128K 18.7 GB

GPT OSS 20B

11.6 GB11.6 GB file

32K 12.1 GB · 128K 13.8 GB

IQ4_NL

Gemma 4 26B A4B IT

15.2 GB13.6 GB file

32K 16.1 GB · 128K 19.4 GB

GPT OSS 20B

12.1 GB11.6 GB file

32K 12.6 GB · 128K 14.3 GB

Q4_1

Gemma 4 26B A4B IT

15.2 GB16.1 GB file

32K 16.1 GB · 128K 19.4 GB

GPT OSS 20B

12.1 GB11.6 GB file

32K 12.6 GB · 128K 14.3 GB

Q4_K_S

Gemma 4 26B A4B IT

15.2 GB16.5 GB file

32K 16.1 GB · 128K 19.4 GB

GPT OSS 20B

12.1 GB11.6 GB file

32K 12.6 GB · 128K 14.3 GB

Q4_K_L

Gemma 4 26B A4B IT

16.5 GB17.2 GB file

32K 17.4 GB · 128K 20.7 GB

GPT OSS 20B

13.2 GB11.9 GB file

32K 13.7 GB · 128K 15.4 GB

Q5_0

Gemma 4 26B A4B IT

16.8 GB17.5 GB file

32K 17.7 GB · 128K 21.0 GB

GPT OSS 20B

not listed

Q5_K_S

Gemma 4 26B A4B IT

18.4 GB18.9 GB file

32K 19.3 GB · 128K 22.6 GB

GPT OSS 20B

14.8 GB11.7 GB file

32K 15.3 GB · 128K 17.0 GB

Q5_K_L

Gemma 4 26B A4B IT

19.4 GB19.5 GB file

32K 20.3 GB · 128K 23.6 GB

GPT OSS 20B

15.5 GB11.9 GB file

32K 16.0 GB · 128K 17.7 GB

BF16

Gemma 4 26B A4B IT

52.3 GB~51.6 GB file

32K 53.2 GB · 128K 56.5 GB

GPT OSS 20B

42.2 GB~41.8 GB file

32K 42.7 GB · 128K 44.4 GB

How to run each

Specifications

Specifications of Gemma 4 26B A4B IT and GPT OSS 20B
SpecGemma 4 26B A4B ITGPT OSS 20B
PublisherGoogleOpenAI
Parameters25.8B20.9B
Active parameters4.4B per token4.2B per token
ArchitectureGemma4ForConditionalGenerationGptOssForCausalLM
Context256K tokens128K tokens
ModalitiesVision, Function-callingChat, Function-calling
LicenseApache 2.0Apache 2.0
ReleasedMar 2026Aug 2025

Frequently Asked Questions

Which is better, Gemma 4 26B A4B IT or GPT OSS 20B?

Gemma 4 26B A4B IT leads llmrun Score 105.4 to 96.2; both fit 24 GB at Q4_K_M. Choose Gemma 4 26B A4B IT for higher llmrun Score and stronger at reasoning; choose GPT OSS 20B for lighter on memory and also does chat.

Which needs less VRAM, Gemma 4 26B A4B IT or GPT OSS 20B?

At Q4_K_M, Gemma 4 26B A4B IT requires 16.1 GB while GPT OSS 20B needs 12.9 GB, making GPT OSS 20B the easier model to run on your local setup.

Which is faster on RTX 4090?

Estimated decode speed on RTX 4090 is about 171 tok/s for Gemma 4 26B A4B IT (Q4_K_M) and 190 tok/s for GPT OSS 20B (Q4_K_M), so GPT OSS 20B is faster. Estimates, ±20%.

Can I run Gemma 4 26B A4B IT or GPT OSS 20B on a 16 GB GPU?

Gemma 4 26B A4B IT fits at Q4_1; GPT OSS 20B fits at Q4_K_M.

Which has a longer context window, Gemma 4 26B A4B IT or GPT OSS 20B?

Gemma 4 26B A4B IT supports 262,144 tokens and GPT OSS 20B supports 131,072 tokens.

Gemma 4 26B A4B IT vs GPT OSS 20B — VRAM & GPU Comparison | llmrun