You can do a lot with the Chinese open source models.

Not in my experience (yes, sample ~= 1), where they are subpar in terms of cybersecurity (e.g. source code review). This is one of my reports using GPT 5.6 (High), GLM 5.3 High, Fable 5.1 [1]:

> On the full benchmark of 31 audited revisions, GPT-5.6 Sol High reached a weighted recall of 38.27% and GLM-5.3 High reached 30.86%. On a common subset of 12 projects, Claude Fable 5.1 High reached 65.70%, against 42.98% for GPT and 35.54% for GLM.

[1] https://github.com/CoinFabrik/soroban-audit-harness/blob/mai...

▲

_rs 3 hours ago | parent | prev [-]

That assumes you have the hardware capable of running inference on them at a reasonable speed

▲

noosphr 3 hours ago | parent | next [-]

Or $10 on open router.

▲

ipsum2 3 hours ago | parent | prev [-]

Or it assumes you have $5 to spend on the tokens for inference. Not everyone has the money.

▲

r_lee 2 hours ago | parent [-]

it takes a lot more than that....

▲

bhelkey 2 hours ago | parent [-]

On OpenRouter, GLM 5.3 Flash costs $0.0285 for 1M input tokens and $0.0923 for 1M output tokens. [1]

[1] https://openrouter.ai/z-ai/glm-5.3-flash

	▲	r_lee an hour ago \| parent [-]
		if you wanna have the same kind of capability as these researchers where you have tight loops calling cli stuff you won't get by with $5 at all, especially for major bugs if you have a way to turn $5 easily into exploitable vulns, let me know cause that'd be a first for me