whatmodel.sh / s1.1 32B
s1.1 32B
↗ share
+ contribute
s1.1 32B is an open-weights language model from Simple Scaling, released February 2025. Its published size is 32.76B parameters and a 33k-token context window. Licence: apache-2.0.
Catalogue identity notes
Maker repo simplescaling/s1.1-32B @7b641db (s1 project, arXiv 2501.19393, Stanford/Hashimoto group): licence tag apache-2.0; successor of s1-32B fine-tuned from Qwen/Qwen2.5-32B-Instruct on s1K-1.1 (DeepSeek-R1 traces). 32,763,876,352 parameters from the Hugging Face safetensors metadata, stored as F32; dense, so active = total. Download 131.06 GB = sum of the safetensors shards at the pinned revision (float32 weights; vLLM's default dtype serves them at half precision). Released 2025-02: month the repo was created (createdAt 2025-02-08). Provider string 'Simple Scaling' is new (HF org simplescaling); no maker logo yet. Distinct from s1-32B (not catalogued). Why proposed: Lands in both Arena-Hard v2 README runs and in the OpenThinker2-32B card runs (as an independent comparator): 6 results.
$ how it compares
The LLMs ranked either side of s1.1 32B, out of 828 with a score.
| # | model | score | where |
|---|---|---|---|
| 357 |  Qwen2.5-VL 32B Instruct | 53±2 | local |
| 358 |  DeepSeek-Coder-V2 Instruct (0614) | 53±1 | local |
| 359 | s1.1 32B | 53±1 | local |
| 360 | DeepSeek-Coder-V2 Instruct (0724) | 53±1 | local |
| 361 |  Olmo 3 7B Think | 53±1 | local |
$ what it's good at
coding▁▁▁▃▆██▇▆▄▃▁ #207 of 458 * · 4.0% of its score · 1 test
Beats Qwen3.5 4B, Qwen3-VL 32B Instruct, Solar Open 100B (102B-A12B) here, though they rank higher overall. Trails QwQ-32B-Preview, Ministral 3 3B Reasoning 2512, DeepSeek-Coder-V2 Instruct (0724) here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LiveCodeBench v2 | 65.5% ✎ | ▄··█·▄···▄▄█ beats 4/7 | 4.0% |
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #241 of 393 * · 14% of its score · 2 tests
Beats Mistral Small 4 (119B-A6.5B), Ministral 3 14B Reasoning 2512, Qwen2.5 Coder 32B Instruct here, though they rank higher overall. Trails Flan-T5 XXL, Granite 4.0 H 1B, LFM2.5 1.2B Instruct here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| Arena-Hard v2.0 hard prompts (style control, Gemini-2.5 judge) | 22.3% ✎ | ▂·▃▅▅·▃█▆▂▂▂ beats 8/22 | 12% |
| Arena-Hard v2.0 creative writing (GPT-4.1 + Gemini-2.5 judges) | 8.2% ✎ | ▆▂▂▆▂▂▃█▂▃·▂ beats 0/22 | 2.6% |
math▁▂▃▅▆▅▄█▄▁▁▁ #188 of 398 · 52% of its score · 3 tests
Beats MiniMax-M2.1, DeepSeek V3 0324, Qwen3 235B-A22B (Apr 2025, hybrid) here, though they rank higher overall. Trails Phi-4-mini-reasoning 3.8B, DeepSeek R1 Distill Qwen 7B, OpenThinker3 7B here, though they rank lower overall.
other tests1 not in a task · 29% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| GPQA Diamond +1 | 62.0% ✎ | ▁▁▁▁▂▄▅▅███▄ beats 118/310 | 29% |
$ where to run it
llama.cpp · recommended
version
- Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.macOS / Linux
curl -LsSf https://llama.app/install.sh | shWindowsirm https://llama.app/install.ps1 | iex - Run s1.1 32B · needs ~19.9 GB · from huggingface.co/bartowskichat in the terminal
llama cli -hf bartowski/simplescaling_s1.1-32B-GGUF:Q4_K_M - Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf bartowski/simplescaling_s1.1-32B-GGUF:Q4_K_M
LM Studio
version
- Install LM Studio · Open the app once before using the lms command.
- Run s1.1 32B · needs ~19.9 GB · from huggingface.co/bartowski
- Or download it in the terminal
lms get https://huggingface.co/bartowski/simplescaling_s1.1-32B-GGUF@Q4_K_M -y - Then run lms ls to see its name, and lms chat followed by that name
Ollama
version
- Install Ollama
download it (macOS / Windows) ↗
Linuxcurl -fsSL https://ollama.com/install.sh | sh - Run s1.1 32B · needs ~19.9 GB · from huggingface.co/bartowskichat in the terminal
ollama run hf.co/bartowski/simplescaling_s1.1-32B-GGUF:Q4_K_M
MLX (Mac)
version
- Install MLX (Mac) · Apple silicon Macs only, text only.with uv
uv tool install mlx-lmorpip install mlx-lm - Run s1.1 32B · needs ~19.4 GB, Mac only · from huggingface.co/mlx-communitychat in the terminal
mlx_lm.chat --model mlx-community/s1.1-32B-4bit - Or start an OpenAI-compatible server at http://localhost:8080/v1
mlx_lm.server --model mlx-community/s1.1-32B-4bit
Other ways to run it
- Run it as published · needs ~131.06 GB, NVIDIA only · from huggingface.co/simplescaling (official)
vllm serve simplescaling/s1.1-32B
Official weights and documentation: huggingface.co/simplescaling ↗
$ how the score was worked out · open-thoughts/OpenThinker2-32… 51%, open-thoughts/OpenThinker-32B… 34%, Arena-Hard-Auto v2.0-Preview… 14%
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 open-thoughts/OpenThinker2-32B model card | 4 | 51% |
| open-thoughts/OpenThinker-32B model card | 4 | 34% |
 Arena-Hard-Auto v2.0-Preview leaderboard (repository README) | 2 | 14% |
| prior · pull toward the middle | 1% | |
| weighted average | = 53.0 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json