# StepFun Step 3.5 Flash

40 Hacker News commenters recommended StepFun Step 3.5 Flash and 13 criticized it, November 2025 to November 2026. It was recommended in 9 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Fast, cheap open-weight model that people find capable for coding, tool calling and personal assistant tasks. Common caveat: poor token efficiency and occasional hallucinations.

## Why people recommend it

- Low cost per token (5 comments): "StepFun is a good enough model for tool calling that so far has been incredibly cheap that I'm happy with it." ([source](https://news.ycombinator.com/item?id=47787076)) "Right now I'm really jazzed about how cheap StepFun 3.5 Flash for these basic personal assistant tasks" ([source](https://news.ycombinator.com/item?id=47725899)) "Which isn't super suprising as StepFun is ~about 5% the price of Sonnet." ([source](https://news.ycombinator.com/item?id=47603300))
- Fast generation speed (4 comments): "being able to boost Step 3.7 Flash up to 40tk/s on my Spark-alike with proper triple head MTP was the thing that made it superior to Qwen 3.6 27B in wall clock time" ([source](https://news.ycombinator.com/item?id=48692245)) "The official Q4_K_S gguf is quite good and has very good 35 tps generation on a M1 mac studio." ([source](https://news.ycombinator.com/item?id=48322495)) "Strong performance, with speed as the highlight—blazing fast, peaking at 350 tokens/s." ([source](https://news.ycombinator.com/item?id=46859600))
- Runs on personal hardware (4 comments): "What’s wild about Step-3.5-Flash isn’t just the quality, it’s how close it is to fitting on personal hardware." ([source](https://news.ycombinator.com/item?id=46972722)) "The official Q4_K_S gguf is quite good and has very good 35 tps generation on a M1 mac studio." ([source](https://news.ycombinator.com/item?id=48322495)) "surprisingly happy to chat about Chinese banned topics for example when run locally" ([source](https://news.ycombinator.com/item?id=48991612))
- Strong coding and instruction following (4 comments): "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." ([source](https://news.ycombinator.com/item?id=48548045)) "conformance to instructions and output quality was better than most of the leading models I've tested" ([source](https://news.ycombinator.com/item?id=47075448)) "For one-shot coding tasks, I found Step 3.5 Flash to be stronger even than Qwen 3.5 397B." ([source](https://news.ycombinator.com/item?id=47612928))

## What people warn about

- Verbose output uses many tokens (4 comments): "it's an amazingly capable model all things considered, but it's token efficiency is so bad" ([source](https://news.ycombinator.com/item?id=48216325)) "token efficiency is so bad that it gets out performed by most others wall-clock time" ([source](https://news.ycombinator.com/item?id=48215398)) "The output was fast, but because it wrote so I had to wait longer." ([source](https://news.ycombinator.com/item?id=47072403))
- Hallucinates and makes mistakes (3 comments): "it hallucinates during a lot of tasks, and the output quality isn't something to write home about." ([source](https://news.ycombinator.com/item?id=49159530)) "it makes some grave mistakes (had some Hanzi in the output once and typos in words)" ([source](https://news.ycombinator.com/item?id=47604963)) "Hallucinates like crazy. use with caution." ([source](https://news.ycombinator.com/item?id=47070128))
- Poor name and website (3 comments): "you really can't see why they wouldn't reach mass market with this name" ([source](https://news.ycombinator.com/item?id=48337606)) "It's hard to take it seriously due to its name and crappy website" ([source](https://news.ycombinator.com/item?id=48324938)) "Have to use browser translation which distorts the website" ([source](https://news.ycombinator.com/item?id=48331536))

## Alternatives hackers recommend

- [Claude Opus](https://hackerapproved.com/p/claude-opus.md): 3331 recommend
- [OpenAI GPT](https://hackerapproved.com/p/gpt-5-5.md): 1205 recommend
- [Claude Sonnet](https://hackerapproved.com/p/claude-sonnet.md): 854 recommend
- [DeepSeek V4 Flash](https://hackerapproved.com/p/deepseek-v4-flash.md): 828 recommend
- [Fable](https://hackerapproved.com/p/fable.md): 823 recommend

## Cite as

Hacker Approved, "StepFun Step 3.5 Flash", https://hackerapproved.com/p/step-3-5-flash/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).