← Best models

Best model for instruction following

Models ranked by their instruction following score, 0 to 100, as on the category page. Only ranked models with enough scored items are included, and models are grouped by total parameters.

Read the memory figure as an estimate, not a guarantee. The memory a model needs is estimated from its file: the weights plus about 20% for the context of a single conversation. It is not a measurement, and a long conversation or several at once needs more. A model with no file size on record is not counted as fitting any amount of memory. Speeds, and the GPU memory shown on the speed page, were measured on this benchmark's server, not on yours. That server kept 1 to 8 answers in flight at a time and loaded a context window of 16,384 tokens for most models; 42 of the models ranked ran at a shorter context. Speeds were measured on one laptop and are comparable only between models on the same backend. See the settings and the memory method.

All sizes

Gemma 4 12B IT leads here with 100.0 on a scale of 0 to 100; its Overall is 98.6 (95% interval 97.0–99.8).

All sizes: the top 5 of 227.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 Gemma 4 12B IT 100.030 of 30 items scored 98.697.0–99.8 1–12 8.1ROCm 7.1 8.9 9–16B
1 gemma-4-26B-A4B-it 100.030 of 30 items scored 97.895.8–99.3 1–22 15.7ROCm 7.1 21.6 16–35B
1 gemma-4-31B-it 100.030 of 30 items scored 99.398.3–100.0 1–5 3.8ROCm 7.1 22.4 16–35B
1 gpt-oss-120b 100.030 of 30 items scored 98.696.8–100.0 1–14 11.0Vulkan 78.4 > 35B
1 Muse Glimmer 30B 100.030 of 30 items scored 97.996.0–99.3 1–23 3.6ROCm 7.1 21.8 16–35B

Tiny (under 3B)

Qwen3 VL 2B Instruct leads its size class with 89.8 on a scale of 0 to 100; its Overall is 77.4 (95% interval 72.5–82.1).

Tiny (under 3B): the top 5 of 49.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 Qwen3 VL 2B Instruct 89.830 of 30 items scored 77.472.5–82.1 68–137 26.5ROCm 7.1 2.3 Tiny (under 3B)
2 Lfm3 VL 2B 88.130 of 30 items scored 72.267.2–77.0 89–158 41.9ROCm 7.1 1.8 Tiny (under 3B)
3 Qwen3 1.7B 86.830 of 30 items scored 76.071.3–80.9 77–141 38.5ROCm 7.1 1.7 Tiny (under 3B)
4 Granite4 1B H 85.130 of 30 items scored 66.861.4–72.1 119–162 24.9ROCm 7.1 1.9 Tiny (under 3B)
5 Gemma2 2B Instruct 84.630 of 30 items scored 68.262.9–73.3 112–160 27.6ROCm 7.1 2.0 Tiny (under 3B)

3–4B

granite-4.1-3b leads its size class with 93.9 on a scale of 0 to 100; its Overall is 80.7 (95% interval 75.9–85.3).

3–4B: the top 5 of 26.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 granite-4.1-3b 93.930 of 30 items scored 80.775.9–85.3 66–129 20.1ROCm 7.1 2.5 3–4B
2 Llama-3.2-3B-Instruct 91.730 of 30 items scored 75.771.0–80.4 84–150 20.7ROCm 7.1 2.4 3–4B
3 Granite4.2 3B 91.230 of 30 items scored 78.273.7–83.1 72–140 24.3ROCm 7.1 2.7 3–4B
4 Granite4 Micro H 90.030 of 30 items scored 76.672.0–81.4 72–139 13.4ROCm 7.1 2.3 3–4B
5 Granite4 3B 87.330 of 30 items scored 74.169.0–79.2 85–145 71.3ROCm 7.1 2.5 3–4B

4–9B

Nemotron 3.5 Bolt E4B leads its size class with 100.0 on a scale of 0 to 100; its Overall is 90.3 (95% interval 86.3–93.9).

4–9B: the top 5 of 69.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 Nemotron 3.5 Bolt E4B 100.030 of 30 items scored 90.386.3–93.9 8–82 9.8ROCm 7.1 8.6 4–9B
2 Qwen3-VL-8B-Instruct 98.530 of 30 items scored 94.992.4–97.1 4–48 9.4ROCm 7.1 7.4 4–9B
3 Olmo-3-7B-Think 98.330 of 30 items scored 77.773.1–82.5 68–146 13.5ROCm 7.1 5.4 4–9B
4 Gemma4 E4B IT 97.230 of 30 items scored 95.593.0–97.7 2–51 13.2ROCm 7.1 7.9 4–9B
5 Olmo-3-7B-Instruct 96.730 of 30 items scored 82.978.8–86.9 52–118 10.2ROCm 7.1 5.4 4–9B

9–16B

Gemma 4 12B IT leads its size class with 100.0 on a scale of 0 to 100; its Overall is 98.6 (95% interval 97.0–99.8).

9–16B: the top 5 of 25.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 Gemma 4 12B IT 100.030 of 30 items scored 98.697.0–99.8 1–12 8.1ROCm 7.1 8.9 9–16B
2 Qwen3.5-9B 94.130 of 30 items scored 92.089.0–95.1 11–69 8.4ROCm 7.1 6.8 9–16B
3 Qwen3-14B 93.630 of 30 items scored 91.788.5–94.7 11–71 5.0ROCm 7.1 11.1 9–16B
4 gemma-2-9b-it 92.630 of 30 items scored 79.975.5–84.2 69–131 7.6ROCm 7.1 6.9 9–16B
5 Qwen2.5-14B-Instruct 91.330 of 30 items scored 91.388.0–94.2 11–72 21.7ROCm 7.1 10.8 9–16B

16–35B

gemma-4-26B-A4B-it leads its size class with 100.0 on a scale of 0 to 100; its Overall is 97.8 (95% interval 95.8–99.3).

16–35B: the top 5 of 49.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 gemma-4-26B-A4B-it 100.030 of 30 items scored 97.895.8–99.3 1–22 15.7ROCm 7.1 21.6 16–35B
1 gemma-4-31B-it 100.030 of 30 items scored 99.398.3–100.0 1–5 3.8ROCm 7.1 22.4 16–35B
1 Muse Glimmer 30B 100.030 of 30 items scored 97.996.0–99.3 1–23 3.6ROCm 7.1 21.8 16–35B
1 Qwen3.5-27B 100.030 of 30 items scored 96.293.8–98.3 3–35 2.7ROCm 7.1 19.9 16–35B
5 Qwen3.8-27B 98.930 of 30 items scored 95.492.9–97.8 2–53 3.5Vulkan 20.2 16–35B

> 35B

gpt-oss-120b leads its size class with 100.0 on a scale of 0 to 100; its Overall is 98.6 (95% interval 96.8–100.0).

> 35B: the top 5 of 8.
Place Model Instruction following (0 to 100) Overall Rank range Decode (tok/s) Memory needed (GB, est.) Size class
1 gpt-oss-120b 100.030 of 30 items scored 98.696.8–100.0 1–14 11.0Vulkan 78.4 > 35B
2 Llama 4 Scout 17B-16E Instruct 98.730 of 30 items scored 93.690.7–96.1 4–65 17.0Vulkan 80.9 > 35B
3 Llama-3.3-70B-Instruct 97.530 of 30 items scored 94.391.3–96.7 4–63 1.4Vulkan 51.0 > 35B
4 Flint 40B A2B 96.630 of 30 items scored 89.085.1–92.6 14–90 25.9ROCm 7.1 27.6 > 35B
5 Qwen3-Coder-Nextcode model 95.230 of 30 items scored 96.494.2–98.3 2–42 18.1Vulkan 62.1 > 35B

A higher place is not always a clear win. When two models have overlapping rank ranges, this battery cannot tell them apart; see how sure a score is. A dash means not measured or not scored, which is different from 0. Speeds were measured on one laptop and are comparable only between models on the same backend.

More: all the best-model pages, the hardware picker, speed and memory for every model.