Step Flash
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
Why people recommend it
“I use Step 3.7 Flash locally, models are good enough for so many coding tasks even at the lower end!”
“I'm still running Step 3.7 Flash at ~40tk/s decode, 1000tk/s+ prefill on mine and its both very capable and fast enough”
Show 3 more
“Step 3.7 Flash at IQ4_XS beats Qwen 3.6 27B despite the latter having MTP, on all of my actual coding task evaluations in real codebases.”
“I’ve been using this model for a while, and it’s very fast.”
Caveats and dissent
“it hallucinates during a lot of tasks, and the output quality isn’t something to write home about”
Often compared with
Who recommends it in their own words
Threads where it came up
- Is Hermes Agent unusable without a model router? 1 mention
- GLM-5.2 is the step change for open agents 1 mention
- Running local models is good now 1 mention
- Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding? 1 mention
- RTX 5080 and RTX 3090 Setup: 80 Tok/s on Qwen 3.6 27B Q8 1 mention
- Step 3.7 Flash 1 mention
Also written as Step 3.5 Flash, Step-3.5-Flash. Data: step-flash.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 8
“it hallucinates during a lot of tasks, and the output quality isn’t something to write home about”
“I use Step 3.7 Flash locally, models are good enough for so many coding tasks even at the lower end!”
“I'm still running Step 3.7 Flash at ~40tk/s decode, 1000tk/s+ prefill on mine and its both very capable and fast enough”
“Step 3.7 Flash at IQ4_XS beats Qwen 3.6 27B despite the latter having MTP, on all of my actual coding task evaluations in real codebases.”
“my local LLM (Step 3.7 Flash)”
“So happy 3.7 came out!”
“I am curious about https://huggingface.co/stepfun-ai/Step-3.5-Flash given that it does parallel token generation.”
“I’ve been using this model for a while, and it’s very fast.”