[whatmodel.sh](../../README.md) / Qwen3 1.7B
# Qwen3 1.7B
Alibaba · released 2025-04 · 2.03B params · apache-2.0
**score 44.5** ±0.5 · #486 of 732 · from 73 benchmark results across 17 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Qwen3 1.7B among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃●█▄▅▅▄▄▂▁ #247 of 313 4.5% 4
coding ▁▁▁▃●▇█▇▅▄▃▁ #372 of 469* 2.4% 2
chat & writing ▁▁▁▃▆▄●█▆▇▇▃ #280 of 402 3.0% 2
math ▁▁▃▅●▆▄█▄▂▁▁ #258 of 431 30% 11
agent tasks · #247 of 313 · 4.5% of its score
Beats [o3-mini](o3-mini.md), [Llama 4 Maverick (402B-A17B)](llama-4-maverick.md), [DeepSeek R1 0528 Qwen3 8B](deepseek-r1-0528-qwen3-8b.md) here, though they rank higher overall. Trails [Granite 4.0 1B](granite-4-0-1b.md), [SmolLM3 3B](smollm3-3b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
BFCL v3 54.7% ▁▁▁▁▁▃▅●▅▄▁▁ beats 56/114 medium · avg 59% 2.2% blunt · at its level · several runs
BFCL V4 (December 2025 archive) 28.4% ▁▁▁▂█●▅▅▆▂▂▂ beats 27/77 hard · avg 31% 1.0% typical
EEGAgentBench overall (tool-based… 23.3% ●·▁▁··▇▇█▂·▁ beats 1/25 hard · avg 28% 0.7% blunt
BFCL (version not stated) 88.8% ▄▆▄▂·█▆·●··▂ beats 15/17 very easy · avg 88% 0.5% blunt · small run
coding · #372 of 469 · 2.4% of its score
Beats [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md), [Qwen3-VL 2B Thinking](qwen3-vl-2b-thinking.md), [DeepSeek R1 Distill Qwen 7B](deepseek-r1-distill-qwen-7b.md) here, though they rank higher overall. Trails [Qwen2.5 1.5B Instruct](qwen2-5-1-5b.md), [Code Tulu 2 13B](code-tulu-2-13b.md), [Phi-3-mini-128k-instruct (Apr 2024)](phi-3-mini-128k-instruct.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LiveCodeBench v6 (window unspecif… 31.3% ▃●█▅▅▃▅···▃█ beats 2/16 hard · avg 35% 1.7% typical
LiveCodeBench v4 34.4% █·●········█ beats 1/2 medium · avg 45% 0.7% at its level · small run
chat & writing · #280 of 402 · 3.0% of its score
Beats [Ministral 3 14B Reasoning 2512](ministral-3-14b-reasoning.md), [Ministral 3 14B](ministral-3-14b.md), [Ministral 3 8B](ministral-3-8b.md) here, though they rank higher overall. Trails [EXAONE 3.5 2.4B Instruct](exaone-3-5-2-4b.md), [Flan-T5 XXL](flan-t5-xxl.md), [LFM2 1.2B](lfm2-1-2b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
IFEval (aggregation unspecified) 74.0% ▄····█▄···▄● beats 5/6 easy · avg 82% 2.0% several runs
IFEval mean of four measures 71.7% █·█··●·██·██ beats 2/6 very easy · avg 86% 1.1% typical
math · #258 of 431 · 30% of its score
Beats [GPT-4o (2024-11-20)](gpt-4o-2024-11-20.md), [GPT-4 Turbo (2024-04-09)](gpt-4-turbo-2024-04-09.md), [Llama 3.1 405B Instruct](llama-3-1-405b.md) here, though they rank higher overall. Trails [DeepSeek R1 Distill Qwen 1.5B](deepseek-r1-distill-qwen-1-5b.md), [Gemma 7B (base)](gemma-7b-base.md), [CodeGemma 7B (v1.0)](codegemma-7b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
GSM8K 77.9% ●····█··▄·▄█ beats 0/7 very easy · avg 86% 12% sharp · at its level · several runs
AIME 2025 (maker) 36.8% ▃·●▃█··▃▅▃·█ beats 1/12 medium · avg 52% 5.8% sharp · at its level · small run · seve…
MGSM 66.6% █···█▄··▄··● beats 6/6 easy · avg 83% 4.2% sharp · at its level
MATH (subset unspecified) 70.6% ●▄··▄█▄····▄ beats 1/7 easy · avg 80% 2.4% at its level
MGSM (Marco evaluation) 63.6% ▄·▄·●▄▄·▄··█ beats 2/7 easy · avg 69% 1.3% at its level
HMMT February 2025 (maker) 10.2% ●▁▁▃▁▅▅▃██▁▄ beats 0/31 hard · avg 34% 1.0% small run
PolyMath (Marco evaluation) 23.4% ▄··●██··▄··▄ beats 1/7 hard · avg 29% 1.0% typical
PolyMATH (maker) 25.2% ▁···●▁·▇█▄▄▄ beats 1/23 hard · avg 29% 0.8% small run
HMMT November 2025 (maker) 8.9% ●▁····▁▅▃█▇▁ beats 0/23 medium · avg 37% 0.7% at its level · small run
MATH-500 81.9% ▄··█▄·▄···●▄ beats 5/6 easy · avg 77% 0.7% at its level · noisy
OTIS mock AIME 2024–2025 8.1% ▃▃▄●▃▃▇▇█▃▂▅ beats 47/197 very hard · avg 7.2% 0.2% sharp · too hard for it · noisy
other tests (32)
test result vs the others in its run difficulty counts why
MMLU-Pro (maker) 58.1% ▄·●·▄▄▄▄█▄██ beats 1/12 medium · avg 63% 11% at its level · several runs
GPQA Diamond 38.0% ▁▁▁▁▂●▆▄▅▇█▅ beats 34/211 medium · avg 52% 8.1% sharp · small run · several runs
MMLU 62.4% ●···▄▄·▄█··█ beats 0/7 easy · avg 74% 7.5% at its level · several runs
GPQA (subset unspecified) (maker) 27.9% ●····▂·▂▂█▆█ beats 0/14 hard · avg 49% 6.0% sharp · several runs
IFBench (aggregation unspecified) 25.9% ▄·█·●······█ beats 4/6 medium · avg 38% 3.4% at its level · several runs
MMMLU 46.5% ▄···▄·▄▄··▄● beats 5/6 medium · avg 57% 2.8% at its level · several runs
MMLU-ProX-Lite (Marco) 28.6% ●··▄··█··█·▄ beats 1/7 medium · avg 42% 1.7% at its level
BRIDGE clinical text, zero-shot a… 26.3% ▁▁▁▁▁▂●▃▄▆█▃ beats 22/97 hard · avg 31% 1.5% blunt · at its level
Open Persian LLM Leaderboard v2 a… 41.3% ▂▄●▃█▄▃▆▃▅▂▂ beats 15/81 medium · avg 57% 1.4% blunt · at its level
INCLUDE (maker) 51.8% ▂····▂▂●▄▄█▄ beats 7/36 medium · avg 59% 1.3% at its level · small run
C-Eval (maker) 68.1% ▁···▁▁▂▁●▂▃█ beats 10/33 easy · avg 75% 1.3% at its level · small run
MMLU-Redux (maker) 73.9% ▁▁▃●▂▃▄▅▇▅█▄ beats 6/63 easy · avg 81% 1.2% at its level · small run
GlobalMMLU (Marco evaluation) 46.3% ▃··●▃··▅···█ beats 1/7 medium · avg 56% 1.2% at its level
MMMLU (Marco evaluation) 49.0% ▄···●·█·▄·▄█ beats 1/7 medium · avg 59% 1.1% at its level
AGIEval (Marco suite unspecified) 39.6% ●··▄█▄··▄··█ beats 0/7 medium · avg 46% 0.9% blunt · at its level
KINA (Knowledge Index of Noah's A… 18.3% ▁▁●▁▃▄█▆▄▅▂▃ beats 3/39 very hard · avg 21% 0.9% typical
MultiChallenge (maker) 27.2% ▁·●·▁▃▆▅▃▂█▅ beats 1/30 hard · avg 35% 0.9% at its level · small run
SuperGPQA (maker) 31.2% ▂▁●▂▅▆▃▇▇▇█▃ beats 4/46 medium · avg 40% 0.9% at its level · small run
Global PIQA (maker) 63.1% ▃●··▃··▃▃▅█▅ beats 1/11 medium · avg 70% 0.8% at its level · small run
MMLU-ProX (maker) 49.4% ▁·●▁▁▂▁▃▃█▅▄ beats 1/36 medium · avg 57% 0.8% at its level · small run
MAXIFE (maker) 50.7% ▂·●·▂···▆·▄█ beats 1/11 medium · avg 63% 0.8% at its level · small run
Global-MMLU 62.3% ▃●▃·▃···▅·▅█ beats 2/11 medium · avg 70% 0.8% at its level · small run
Open PT-LLM Leaderboard average (… 68.1% ▃▁▁▁▁▁▁▁▂●▅▁ beats 148/276 medium · avg 72% 0.7% at its level · noisy
EuroEval HellaSwag (English), tes… 46.3% ▃·▁▁▂▁●▅█▇▄▁ beats 24/100 medium · avg 67% 0.6% sharp · at its level · noisy
MixEval-Hard 33.9% █····●·····█ beats 1/2 medium · avg 39% 0.6% at its level · small run
LongBench v2 (maker) 26.5% ●·▁·▂▃▁▄▄▃█▄ beats 1/33 very hard · avg 34% 0.5% too hard for it · noisy · small run
NOVA-63 (maker) 40.3% ●▃·▃···▃█▃▃█ beats 0/11 medium · avg 47% 0.5% blunt · at its level · small run
EuroEval Life in the UK (English … 68.4% ▁▁▁▁▁▁▁▂●█▄▃ beats 36/146 easy · avg 82% 0.4% at its level · noisy
GSM-Plus 79.4% ▁··▁▁▁·●▁▄█▁ beats 4/16 easy · avg 83% 0.4% blunt · small run
Vectara hallucination leaderboard… 95.6% ▁▁▂▃▄▄●▆▃▂▂▁ beats 78/140 very easy · avg 94% 0.2% blunt · noisy
WMT24++ (maker) 39.3% ▁·●▁·▁·▁▂▆█▃ beats 1/27 medium · avg 62% 0.1% blunt · at its level · small run
Chess Puzzles (Epoch) 0.0% ●▁▁▁▁▂▃▃▃▂▁▁ beats 33/137 very hard · avg 5.8% 0.0% sharp · too hard for it · noisy
source results share of the score
Marco-MoE: Open Multilingual Mixture-of… 11 23% ▇▇▇
Qwen/Qwen3.5-2B model card (submitted /… 19 17% ▇▇
LiquidAI/LFM2-1.2B model card 7 15% ▇▇
Qwen (submitted / maker-reported) 5 13% ▇▇
HuggingFaceTB/SmolLM3-3B model card 8 9% ▇
LiquidAI/LFM2.5-1.2B-Thinking model card 8 9% ▇
Unsupervised Confidence Calibration for… 1 2% ▇
Epoch AI 3 2% ▇
Qwen/Qwen3.5-0.8B model card (submitted… 1 2% ▇
Berkeley Function Calling Leaderboard r… 2 2% ▇
BRIDGE Medical Leaderboard 1 1% ▇
Open Persian LLM Leaderboard v2 1 1% ▇
EuroEval leaderboards 2 1% ▇
Knowledge Index of Noah's Ark 1 1% ▇
Open Portuguese LLM Leaderboard 1 1% ▇
EEGAgentBench: Benchmarking LLM Agents … 1 1% ▇
Vectara hallucination leaderboard 1 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 44.5
Licences: Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient Upcycling ([CC BY 4.0](../../licenses/cards-final-NOTICE.txt)) · Qwen/Qwen3.5-2B model card ([apache-2.0](../../licenses/Qwen--Qwen3.5-2B.txt)) · LiquidAI/LFM2-1.2B model card ([LFM Open License v1.0](https://huggingface.co/LiquidAI/LFM2-1.2B/blob/40f3da0d0164913923aee9462c23077868b816a3/LICENSE)) · HuggingFaceTB/SmolLM3-3B model card ([Apache-2.0](../../licenses/HuggingFaceTB--SmolLM3-3B.txt)) · LiquidAI/LFM2.5-1.2B-Thinking model card ([LFM Open License v1.0](../../licenses/cards-gap-NOTICE.txt)) · Unsupervised Confidence Calibration for Reasoning LLMs from a Single Generation ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Qwen/Qwen3.5-0.8B model card ([apache-2.0](../../licenses/Qwen--Qwen3.5-0.8B.txt)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · Open Persian LLM Leaderboard v2 ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE)) · Knowledge Index of Noah's Ark ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · EEGAgentBench: Benchmarking LLM Agents on Short- and Long-Horizon EEG Analysis ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Qwen3 1.7B (Q4_K_M, needs ~1.3 GB, from [huggingface.co/ggml-org](https://huggingface.co/ggml-org/Qwen3-1.7B-GGUF/tree/daeb8e2d528a760970442092f6bf1e55c3b659eb))
```sh
llama cli -hf ggml-org/Qwen3-1.7B-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf ggml-org/Qwen3-1.7B-GGUF:Q4_K_M
```
Other versions: Q8_0 (~2.2 GB) `llama cli -hf ggml-org/Qwen3-1.7B-GGUF:Q8_0`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Qwen3 1.7B (Q4_K_M, needs ~1.3 GB, from [huggingface.co/ggml-org](https://huggingface.co/ggml-org/Qwen3-1.7B-GGUF/tree/daeb8e2d528a760970442092f6bf1e55c3b659eb))
```sh
lms get https://huggingface.co/ggml-org/Qwen3-1.7B-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~2.2 GB) `lms get https://huggingface.co/ggml-org/Qwen3-1.7B-GGUF@Q8_0 -y` · 4bit (~1 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/Qwen3-1.7B-MLX-4bit -y` · 8bit (~1.8 GB, Mac only) `lms get https://huggingface.co/lmstudio-community/Qwen3-1.7B-MLX-8bit -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Qwen3 1.7B (Q4_K_M, needs ~1.72 GB, from [ollama.com](https://ollama.com/library/qwen3/tags))
```sh
ollama run qwen3:1.7b
```
Other versions: Q8_0 (~2.94 GB) `ollama run qwen3:1.7b-q8_0`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run Qwen3 1.7B (4bit, needs ~1 GB, Mac only, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Qwen3-1.7B-MLX-4bit/tree/46a51038b9871ba2ad249422905f1defe1f289f5))
```sh
mlx_lm.chat --model lmstudio-community/Qwen3-1.7B-MLX-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model lmstudio-community/Qwen3-1.7B-MLX-4bit
```
Other versions: 8bit (~1.8 GB, Mac only) `mlx_lm.chat --model lmstudio-community/Qwen3-1.7B-MLX-8bit`
Model repository and documentation: [huggingface.co/Qwen](https://huggingface.co/Qwen/Qwen3-1.7B)
## Published speed measurements
Reference results for the hardware, engine and settings below. These are not verified speeds of the downloads above or predictions for your machine.
hardware precision speed engine and settings
NVIDIA GeForce RTX 3050 Ti … q4_k_m 95.98 tok/s ollama (version not recorded) · l…
Each speed links to its source, which has the full settings.
## Quantization evidence
How much of the full-precision result a quantized version kept, measured within one study. Rows from the same study share a table and aren't independent. They don't describe every download of that precision.
precision test full → quantized kept
AutoRound 2-bit (group size 3… WMT24++ X->English COMET (wmt22-c… 0.729 → 0.401 55.0%
AutoRound 2-bit (group size 3… WMT24++ English->X COMET (wmt22-c… 0.554 → 0.302 54.4%
AutoRound 4-bit (group size 1… WMT24++ X->English COMET (wmt22-c… 0.729 → 0.718 98.5%
AutoRound 4-bit (group size 1… WMT24++ English->X COMET (wmt22-c… 0.554 → 0.531 95.8%
AWQ 4-bit (AutoAWQ, zero poin… WMT24++ X->English COMET (wmt22-c… 0.729 → 0.715 98.1%
AWQ 4-bit (AutoAWQ, zero poin… WMT24++ English->X COMET (wmt22-c… 0.554 → 0.523 94.4%
bitsandbytes NF4 (Transformer… WMT24++ X->English COMET (wmt22-c… 0.729 → 0.721 98.9%
bitsandbytes NF4 (Transformer… WMT24++ English->X COMET (wmt22-c… 0.554 → 0.529 95.5%
FP8 dynamic (weights FP8 per-… GPQA Diamond — published row 'GPQ… 0.384 → 0.389 101.3%
FP8 dynamic (weights FP8 per-… AIME 2025 — published row 'AIME 2… 0.323 → 0.323 100.0%
FP8 dynamic (weights FP8 per-… OpenLLM v1 average of 6 tasks (MM… 0.498 → 0.491 98.6%
FP8 dynamic (weights FP8 per-… MMLU — published row 'MMLU (5-sho… 0.568 → 0.56 98.6%
FP8 dynamic (weights FP8 per-… OpenLLM v2 average of 6 tasks (MM… 0.243 → 0.233 95.8%
FP8 dynamic (weights FP8 per-… GSM8K — published row 'GSM-8K (5-… 0.437 → 0.415 95.0%
FP8 dynamic (weights FP8 per-… AIME 2024 — published row 'AIME 2… 0.44 → 0.401 91.2%
GGUF Q2_K with imatrix (llama… WMT24++ X->English COMET (wmt22-c… 0.729 → 0.63 86.5%
GGUF Q2_K with imatrix (llama… WMT24++ English->X COMET (wmt22-c… 0.554 → 0.449 81.0%
GGUF Q4_K_M with imatrix (lla… WMT24++ X->English COMET (wmt22-c… 0.729 → 0.725 99.5%
GGUF Q4_K_M with imatrix (lla… WMT24++ English->X COMET (wmt22-c… 0.554 → 0.547 98.8%
GPTQ INT8 (w8, group_size 128… MMLU-Redux, thinking mode (Qwen-r… 0.739 → 0.749 101.4%
GPTQ INT8 (w8, group_size 128… LiveBench 2024-11-25, non-thinkin… 0.356 → 0.355 99.7%
GPTQ INT8 (w8, group_size 128… GPQA (subset not stated in card),… 0.286 → 0.283 99.0%
GPTQ INT8 (w8, group_size 128… MMLU-Redux, non-thinking mode (Qw… 0.644 → 0.63 97.8%
GPTQ INT8 (w8, group_size 128… GPQA (subset not stated in card),… 0.401 → 0.391 97.5%
GPTQ INT8 (w8, group_size 128… LiveBench 2024-11-25, thinking mo… 0.511 → 0.498 97.5%
INT4 W4A16 (weight-only INT4,… MMLU — published row 'MMLU (5-sho… 0.568 → 0.551 97.0%
INT4 W4A16 (weight-only INT4,… OpenLLM v1 average of 6 tasks (MM… 0.498 → 0.468 94.0%
INT4 W4A16 (weight-only INT4,… OpenLLM v2 average of 6 tasks (MM… 0.243 → 0.224 92.1%
INT4 W4A16 (weight-only INT4,… GPQA Diamond — published row 'GPQ… 0.384 → 0.348 90.8%
INT4 W4A16 (weight-only INT4,… AIME 2025 — published row 'AIME 2… 0.323 → 0.282 87.4%
INT4 W4A16 (weight-only INT4,… LiveCodeBench — published row 'Li… 0.334 → 0.264 78.9%
INT4 W4A16 (weight-only INT4,… AIME 2024 — published row 'AIME 2… 0.44 → 0.321 73.0%
INT4 W4A16 (weight-only INT4,… GSM8K — published row 'GSM-8K (5-… 0.437 → 0.306 70.1%
Each kept value links to its source.
[interactive version](https://whatmodel.sh/model/qwen3-1-7b) · [JSON](https://whatmodel.sh/api?format=json&all=1)