Skip to content
Weights Not Walls
Menu

stepfun-ai

Step-3.7-Flash

Reviewedaccepts images
Size
201B parameterscounted from the safetensors header on the Hub
Licence
unknown, Apache License 2.0Our reading, check the licence: You can use, change and redistribute the model, including commercially. Keep the licence and any notices, and say which files you changed; you do not have to publish your changes. It includes a patent licence that ends for anyone who sues over patents in the work.
On the Hub since
23 May 2026the date the repo was created
Context length
256k tokensfrom the model's config
Kind
Base model (no base model declared on the Hub)Step3p7ForConditionalGeneration
On Hugging Face
stepfun-ai/Step-3.7-Flash9.6k downloads in the last 30 days, 463 likes

Scores

1 independent result on 1 eval set
CritPtReported

Epoch's name: step-3.7-flash. Best closed model on this set: gemini-4-carbon_high, 36.0% (Google DeepMind). All 271 results

Claimed by the maker Claimed

The model card has no claimed scores in a machine-readable form (the card's model-index), so none are shown. The maker may give scores in the card text: read the card.

Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.

Independent scores from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Each row links its own source. What the badges mean.

Run it yourself

file sizes read from the Hub 19 Nov 2026

Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 3.6 GB, worked out from the model's config (stepfun-ai/Step-3.7-Flash), at 16-bit precision. Choose a machine above to see a verdict for each format. Method

GGUF for llama.cpp, Ollama, LM Studio and others

stepfun-ai/Step-3.7-Flash-GGUF (the maker's own conversion), 3.7k downloads. Sizes are the real files on the Hub, all parts added up. For image input you also need the projector file mmproj-step3.7-flash-f16.gguf (3.7 GB), not counted below.

FormatSizeMemory neededGet it
IQ3_XXS70.6 GBfile size, 2 parts74.6 GB70.6 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:IQ3_XXS
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "IQ3_XXS/Step-3.7-flash-IQ3_XXS-*-of-00002.gguf" --local-dir Step-3.7-Flash-GGUF
Q3_K_M87.4 GBfile size, 3 parts91.4 GB87.4 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:Q3_K_M
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "Q3_K_M/Step-3.7-flash-Q3_K_M-*-of-00003.gguf" --local-dir Step-3.7-Flash-GGUF
Q3_K_L95.5 GBfile size, 3 parts99.5 GB95.5 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:Q3_K_L
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "Q3_K_L/Step-3.7-flash-Q3_K_L-*-of-00003.gguf" --local-dir Step-3.7-Flash-GGUF
IQ4_XS97.8 GBfile size, 3 parts101.8 GB97.8 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:IQ4_XS
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "IQ4_XS/Step-3.7-flash-IQ4_XS-*-of-00003.gguf" --local-dir Step-3.7-Flash-GGUF
Q4_K_S103.8 GBfile size, 3 parts107.9 GB103.8 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:Q4_K_S
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "Q4_K_S/Step-3.7-flash-Q4_K_S-*-of-00003.gguf" --local-dir Step-3.7-Flash-GGUF
Q8_0195.0 GBfile size, 5 parts199.1 GB195.0 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:Q8_0
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "Q8_0/Step-3.7-flash-Q8_0-*-of-00005.gguf" --local-dir Step-3.7-Flash-GGUF
BF16367.0 GBfile size, 9 parts371.0 GB367.0 + 3.6 + 0.50
Commands
llama.cpp
llama-server -hf stepfun-ai/Step-3.7-Flash-GGUF:BF16
Download
hf download stepfun-ai/Step-3.7-Flash-GGUF --include "BF16/Step3.7-flash-bf16-*-of-00009.gguf" --local-dir Step-3.7-Flash-GGUF

MLX for Apple silicon

FormatSizeMemory neededRepo
MLXlemuralabs/Step-3.7-Flash-OptiQ-3.7bpw-mlx99.1 GBfile size103.1 GB99.1 + 3.6 + 0.50Download page
MLX 4bitmlx-community/Step-3.7-Flash-4bit103.2 GBfile size107.3 GB103.2 + 3.6 + 0.50Download page

Original weights size from the safetensors header, for vLLM, SGLang, Transformers

FormatSizeMemory neededRepo
BF16original weights375.1 GBfile size379.1 GB375.1 + 3.6 + 0.50
hf download stepfun-ai/Step-3.7-Flash --local-dir Step-3.7-Flash
Other conversions (7; sizes not read)

The Hub tags 40 quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.

Run it by API

Hugging Face Inference Providers, fetched 19 Nov 2026, 13:58 UTC

4 providers serve this model through the Hugging Face router. Prices are US dollars per million tokens. Precision is what the provider reports; a lower precision than the original weights can change answers. Live first, then cheapest output.

ProviderInOutContextPrecisionTokens/sFirst tokenUptime 24hToolsJSON
deepinfra$0.20$1.15256knot given270.6 s?yesyes
Novita$0.20$1.15256kfp8235.6 s100.0%yesyes
StepFun$0.20$1.15250kfp81053.8 s99.8%yesyes
featherless-ai???not given???not givennot given

Through the router, with a Hugging Face token (the router bills your Hugging Face account). Change the name after the colon to pick another provider.

curl https://router.huggingface.co/v1/chat/completions \ -H "Authorization: Bearer $HF_TOKEN" -H "Content-Type: application/json" \ -d '{"model": "stepfun-ai/Step-3.7-Flash:deepinfra", "messages": [{"role": "user", "content": "Hello"}]}'

Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.

Planned for this page Not built yet

Try it in private chat
A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
Fine-tune it
Fine-tuning a few base models from this page and publishing the result here.
Creators and earnings
Pages for people who publish fine-tunes and adapters of this model, and what they earn.
Fund and bounties
Bounties for new open models, paid from a public fund. How the fund will work.
Contribute compute
How to give compute to open model work, with Akash.

Model data from the Hugging Face Hub 19 Nov 2026, 13:43 UTC. The repo. Found a mistake on this page? Tell us.