// head_to_head
gpt-oss-safeguard-20b vs GLM 5.3 Flash
Published LiveBench scores across all seven categories, live list pricing, context windows, and the measured cost of a point of capability — for both models, side by side.
gpt-oss-safeguard-20b is the cheaper of the two; neither can be ranked on quality here.
gpt-oss-safeguard-20b does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.
openai
gpt-oss-safeguard-20b
- Blended / 1M
- $0.131
- Context
- 131K
- Released
- Oct 29, 2025
- Overall score
- Not evaluated
z-ai
GLM 5.3 Flash
- Blended / 1M
- $0.237
- Context
- 1.3M
- Released
- Aug 26, 2026
- Overall score
- 71.6
Specs and pricing
| Metric | gpt-oss-safeguard-20b | GLM 5.3 Flash |
|---|---|---|
| LiveBench overall Mean of the seven LiveBench category scores, 0–100. Higher is better. | — | 71.6 |
| Cost per point Measured benchmark spend divided by overall score — dollars per point of capability. | — | $0.0161 |
| Blended price / 1M 3:1 input:output mix, the usual shape of production traffic. | $0.131win | $0.237 |
| Input price / 1M | $0.075win | $0.150 |
| Output price / 1M | $0.300win | $0.500 |
| Cached input / 1M Price of an input token served from the prompt cache, where the provider publishes one. | $0.037win | $0.050 |
| Context window | 131K | 1.3Mwin |
| Max output tokens | 66K | 944Kwin |
Benchmarks by category
An overall score averages away the thing you probably care about. These are the seven LiveBench categories scored separately — gpt-oss-safeguard-20b on top, GLM 5.3 Flash below, both out of 100.
What each one costs to run
Per-token prices are hard to feel. These are monthly list costs for both models across five workload shapes, using each provider's published cached-input rate where there is one.
| Workload | gpt-oss-safeguard-20b | GLM 5.3 Flash |
|---|---|---|
| Support chatbot 1.2K in / 400 out × 200K requests | $39.30/mo | $68.80/mo |
| RAG assistant 8K in / 600 out × 100K requests | $63.00/mo | $110.00/mo |
| Coding agent 40K in / 4K out × 20K requests | $63.00/mo | $104.00/mo |
| Document extraction 20K in / 1.5K out × 50K requests | $95.63/mo | $182.50/mo |
| Bulk classification 500 in / 20 out × 5M requests | $198.75/mo | $375.00/mo |
Which should you pick?
You need to fit large documents in one call
GLM 5.3 Flash
Wider context window — 1.3M against 131K.
You are cost-constrained
gpt-oss-safeguard-20b
Cheaper on blended list price at $0.131 per million tokens.
gpt-oss-safeguard-20b vs GLM 5.3 Flash FAQ
Which is better, gpt-oss-safeguard-20b or GLM 5.3 Flash?
gpt-oss-safeguard-20b is the cheaper of the two; neither can be ranked on quality here. gpt-oss-safeguard-20b does not have a published LiveBench run, so this comparison covers price, context and declared capabilities only. A missing score means "not evaluated", not "worse" — the right way to separate these two is an eval on your own workload.
Is gpt-oss-safeguard-20b cheaper than GLM 5.3 Flash?
gpt-oss-safeguard-20b is cheaper. On a 3:1 input:output blend, gpt-oss-safeguard-20b lists at $0.131 per million tokens and GLM 5.3 Flash at $0.237 — gpt-oss-safeguard-20b is 1.8× cheaper. Input and output are priced separately — gpt-oss-safeguard-20b charges $0.075 in and $0.300 out, GLM 5.3 Flash charges $0.150 and $0.500 — so the model that looks cheaper flips depending on how output-heavy your workload is.
Does gpt-oss-safeguard-20b or GLM 5.3 Flash have a bigger context window?
GLM 5.3 Flash has the larger context window: 131K for gpt-oss-safeguard-20b against 1.3M for GLM 5.3 Flash. Note that a window you can fill is not a window you should fill — retrieval quality usually degrades well before the limit, and you pay for every token you put in it.
Do gpt-oss-safeguard-20b and GLM 5.3 Flash support prompt caching?
Both publish a cached-input rate: $0.037 per million for gpt-oss-safeguard-20b and $0.050 for GLM 5.3 Flash, against full input rates of $0.075 and $0.150. On a workload with a long stable prefix — a system prompt, a tool schema, a retrieved corpus — that changes the economics more than the headline price does.
Related comparisons
How these numbers are produced
- Price — provider list price from OpenRouter, refreshed every 15 minutes. “Blended” is a 3:1 input:output mix.
- Scores — LiveBench release 2026-06-25, using their own category map. Each model shows its strongest published run. A blank means “not evaluated”, never “bad”.
- Cost per point — the measured dollars LiveBench spent on the run, divided by the score it earned.
- “Win” — awarded only past a threshold: one full point on a benchmark score, 10% on a price, 25% on a context window. Anything tighter reports as a tie, because effort settings alone move a LiveBench score by more than that.
Published benchmarks rank models on someone else's tasks. Before committing, see LLM & agent evaluation for building an eval on your own.