GLM-5.3 Flash is Z.ai's cheapest model at $0.15/1M input tokens — 89% less than the flagship GLM-5.2. It is also the best capability-per-dollar pick in the lineup.
The shortest way to see the safest default, the lower-cost option, and the specialist pick before you read deeper.
GLM-5.3 Flash is the safest overall answer here when you want the strongest default instead of the lowest list price.
Google: Gemini 2.0 Flash is the lower-cost option to start with when you still need useful output at scale.
GLM-5.2 is the better pick when response speed matters more than maximum reasoning depth.
GLM-5.3 Flash is the lowest-cost Z.ai model: $0.15/1M input, $0.5/1M output.
GLM-5.3 Flash is the best capability-per-dollar pick (budget score 95/100).
GLM-5.2 costs 9x more on input — reserve it for work where quality is the bottleneck.
Choose GLM-5.3 Flash for high-volume, low-stakes tasks like classification, extraction, and drafts.
Choose GLM-5.3 Flash as the everyday default if you want one budget model.
Route only the hardest tasks to GLM-5.2 — a two-tier setup usually cuts spend 60–80%.
Switch the scoring lens to see whether the top answer changes when you care more about cost, speed, or long-document work.
Z.ai / Budget / Aug 27, 2026
Native vision and video, MIT weights, fifteen cents per million.
Ranks models by the broadest mix of coding, writing, research, and long-context usefulness.
You need top-tier reasoning or a published SWE-bench Verified figure — this is a volume model, not a ceiling model.
The fastest way to see where the recommendation shifts when your priority changes.
Native vision and video, MIT weights, fifteen cents per million.
Top open-weights coder — beats GPT-5.5 at a sixth of the cost.
Same price as GLM-5.2, far stronger on agents and security.
Every figure below is the provider's list price or a published capability score — the same numbers the recommendation on this page is built from.
| Model | Input | Output | Est. month | Context | Speed | Coding | Writing | Research |
|---|---|---|---|---|---|---|---|---|
| GLM-5.3 FlashZ.ai | $0.15/1M | $0.50/1M | $2.50 | 1M tokens | Fast | 82 | 76 | 78 |
| GLM-5.2Z.ai | $1.40/1M | $4.40/1M | $23 | 1M tokens | Balanced | 90 | 78 | 80 |
| GLM-5.3Z.ai | $1.40/1M | $4.40/1M | $23 | 1M tokens | Balanced | 91 | 79 | 82 |
Capability scores are out of 100 and reflect our own weighting of published benchmarks and production signals — see how we evaluate models. “Est. month” assumes 10M input and 2M output tokens at list price, with no batch or caching discounts applied, so treat it as a ceiling.
What each one is genuinely good at, where it falls down, and the situations we would steer you away from it — not just the headline score.
A 320B-A18B mixture-of-experts model with native vision and video, MIT-licensed weights and a 1M context — at fifteen cents per million input tokens.
You need top-tier reasoning or a published SWE-bench Verified figure — this is a volume model, not a ceiling model.
The cheapest genuinely multimodal 1M-context model worth using in August 2026. If you are running vision or video at volume and can host weights, nothing at this price is close.
Released August 26, 2026. 320B total parameters, 18B active (320B-A18B MoE). Z.ai lists $0.15/1M input, $0.03/1M cached input and $0.50/1M output, with a launch promotion halving those rates through September 9, 2026. Self-reported against GLM-5.2: DeepSWE 63.4 vs 46.2, AutomationBench 48.8 vs 26.2.
Z.ai's MIT-licensed open-weight flagship — the top open-weights coding model of mid-2026, beating GPT-5.5 on agentic coding benchmarks at roughly a sixth of the cost.
You need frontier reasoning ceiling or launch-day verified benchmarks.
The open-weights coding value king of mid-2026 — GPT-5.5-beating agentic coding at a fraction of the price, with an MIT license. DeepSeek V4-Flash undercuts it on price; GLM-5.2 answers with higher ceiling and two reasoning-effort modes.
Announced June 13, 2026; pay-per-token API live June 16. Two reasoning modes ('thinking' and 'max thinking'). GLM Coding Plan: Lite $18/mo (~$12.60 effective yearly), Pro $72, Max $160.
Z.ai's newest flagship, aimed squarely at software engineering, autonomous agents and cybersecurity — and the first open-weights model to beat Claude Mythos 5 on a security benchmark.
Your procurement process requires a SWE-bench Verified figure, or you need the closed-frontier reasoning ceiling.
The clear upgrade over GLM-5.2 at the same price — take it unless you specifically need a published SWE-bench Verified number to sign off on. The security benchmark lead over Mythos 5 is the genuinely new thing here.
Released August 14, 2026. Z.ai list pricing is $1.40/$4.40, the same rate as GLM-5.2; resellers discount from that list. Also available through the GLM Coding Plan from $18/mo. Reported GPQA Diamond 91.7% and Artificial Analysis Intelligence Index 59.5.
UseRightAI recommendations are based on practical decision factors people actually feel in day-to-day use.
Newsletter
Useful if you care about ranking shifts, pricing changes, or a better recommendation appearing in this decision path.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.
GLM-5.3 Flash at $0.15/1M input and $0.5/1M output tokens. Native vision and video, MIT weights, fifteen cents per million.
GLM-5.3 Flash is the best capability-per-dollar pick in Z.ai's lineup (budget score 95/100). It handles cheap multimodal work at scale on MIT-licensed weights well — step up to GLM-5.2 only where quality visibly falls short.
GLM-5.3 Flash costs $0.15/1M input vs $1.4/1M for GLM-5.2 — a 89% saving on input tokens.
GLM-5.3 Flash — 1M tokens at $0.15/1M input. Context is where budget models are least compromised: you usually lose reasoning depth before you lose window size, so a cheap model is often a perfectly good choice for summarising or extracting from long documents.
No published SWE-bench Verified score. The launch promotion halves these rates only through September 9, 2026 — the $0.15/$0.50 list price applies after that. Avoid it if you need top-tier reasoning or a published SWE-bench Verified figure — this is a volume model, not a ceiling model.
On a moderate workload of 10M input and 2M output tokens, GLM-5.3 Flash runs about $2.50 against $22.80 for GLM-5.2 — a difference of $20.30 a month at the same volume. Output tokens dominate the bill on both, so the length of the responses you generate matters far more than the length of your prompts.
Mixing is almost always cheaper for the same quality. Route high-volume, low-stakes work — classification, extraction, first drafts, routine agent steps — to GLM-5.3 Flash, and reserve GLM-5.2 for the calls where a wrong answer costs real time. Teams that split this way typically cut spend substantially without a quality drop anyone notices, because most tokens in a real workload are not hard problems.