MODEL PROOF
GLM 5.3
TRADE-OFF
A stored alternative has an equal-or-higher score and an equal-or-lower price, with capability losses.
$1.00 / 1M tokens Balanced · 3 tokens in per 1 out
Price under confirmation since 2026-10-12. Upstream lists another price for this model now, or other prices from its sellers. What is shown is the last we published; the new figures are published once a second day shows the same values. · Serving precision differs between offers.
MiMo-V2.6-Pro · Δ score 18.9 points · 46% lower measured price
- 944K → 131K max output
- Independent LMArena score
- 1,471.4 ± 5.23 · 21,797 votes
- Context window
- 1.05M
- Maximum output
- 943.72K
- Input modalities
- text
- Output modalities
- text
- Published input price
- $0.200 / 1M tokens
- Published output price
- $3.40 / 1M tokens
- Pricing kind
- fixed
Inspect complete billing conditions and endpoint terms below.
Price from Sail Research: the cheapest offer that is serving, at standard delivery, with at least 90% of the reference context and at the maker’s precision or better, at its standard rate.
Price under confirmation since 2026-10-12 Upstream lists another price for this model now, or other prices from its sellers. What is shown is the last we published; the new figures are published once a second day shows the same values.
Cheaper right now: $0.645/M at Novita — promotion 70% It does not pass the like-for-like test, so it never sets a rank.
Every price dimension · USD per million tokens
| Input | $0.200/M |
|---|---|
| Output | $3.40/M |
| Cached input | $0.150/M |
| Cache write | Unknown/M |
| Cache write 1h | Unknown/M |
| Reasoning | Unknown/M |
Who sells it
The same weights, different shops. Cheapest is not like-for-like when serving precision differs.
Headline cheapest is a lower precision. Like-for-like at the best declared precision is $0.190/M from AkashML.
41 provider offers · rates, limits and conditions
Rates are USD per million tokens. Endpoint terms can differ even when the quoted price and precision match. A listed parameter is a provider declaration, not a task-success test.
| Seller | Input | Output | Precision | Uptime (1d) |
|---|---|---|---|---|
| Relacerelace Endpoint terms
| $0.030 | $12.00 | undeclared | 99.77% |
| Waferwafer/us Endpoint terms
| $0.070 | $7.00 | undeclared | 99.79% |
| InferenceNetinference-net Endpoint terms
| $0.080 | $4.40 | undeclared | 99.64% |
| Makoramakora/fp4 Endpoint terms
| $0.140 | $4.40 | fp4 | 96.23% |
| Waferwafer Endpoint terms
| $0.150 | $7.00 | undeclared | 99.43% |
| Rekareka Endpoint terms
| $0.170 | $3.00 | undeclared | 99.75% |
| AkashMLakashml/fp8 Endpoint terms
| $0.190 | $4.40 | fp8 | 99.93% |
| Sail Researchsail-research/us Endpoint terms
| $0.200 | $3.40 | fp8 | 99.82% |
| Sail Researchsail-research/fp8 Endpoint terms
| $0.200 | $3.40 | fp8 | 99.93% |
| Morphmorph/fp8 Endpoint terms
| $0.238 | $3.55 | fp8 | 99.08% |
| Novitanovita/fp8 Endpoint terms
| $0.420 | $1.32 | fp8 | 98.26% |
| DeepInfradeepinfra/fp4 Endpoint terms
| $0.563 | $2.50 | fp4 | 96.94% |
| Inceptroninceptron/fp4 Endpoint terms
| $0.600 | $3.39 | fp4 | 98.20% |
| SiliconFlowsiliconflow/fp8 Endpoint terms
| $0.700 | $2.20 | fp8 | 99.75% |
| Phalaphala Endpoint terms
| $0.840 | $2.64 | undeclared | 99.12% |
| DigitalOceandigitalocean Endpoint terms
| $0.910 | $2.86 | undeclared | 99.83% |
| GMICloudgmicloud/fp8 Endpoint terms
| $0.980 | $3.08 | fp8 | 98.80% |
| Decartdecart/fp4 Endpoint terms
| $1.19 | $3.74 | fp4 | 99.90% |
| Alibabaalibaba Endpoint terms
| $1.19 | $3.74 | undeclared | 99.88% |
| Friendlifriendli Endpoint terms
| $1.26 | $3.96 | undeclared | 99.95% |
| Baidubaidu/fp8 Endpoint terms
| $1.40 | $4.40 | fp8 | 99.87% |
| Parasailparasail/fp8 Endpoint terms
| $1.40 | $4.40 | fp8 | 99.70% |
| AtlasCloudatlas-cloud/fp8 Endpoint terms
| $1.40 | $4.40 | fp8 | 99.98% |
| Z.AIz-ai/fp8 Endpoint terms
| $1.40 | $4.40 | fp8 | 99.86% |
| BaseTenbaseten/fp4 Endpoint terms
| $1.40 | $4.40 | fp4 | 98.04% |
| Nebiusnebius/fp4 Endpoint terms
| $1.40 | $4.40 | fp4 | 94.68% |
| Crusoecrusoe/fp4 Endpoint terms
| $1.40 | $4.40 | fp4 | 98.57% |
| BaseTenbaseten/fp4 Endpoint terms
| $1.40 | $4.40 | fp4 | 94.31% |
| Mistralmistral/zdr Endpoint terms
| $1.40 | $4.40 | nvfp4 | 99.61% |
| Mistralmistral/nvfp4 Endpoint terms
| $1.40 | $4.40 | nvfp4 | 99.05% |
| PrimeIntellectprimeintellect Endpoint terms
| $1.40 | $4.40 | undeclared | 99.68% |
| Venicevenice Endpoint terms
| $1.40 | $4.40 | undeclared | 98.63% |
| Togethertogether Endpoint terms
| $1.40 | $4.40 | undeclared | 97.18% |
| Modalmodal Endpoint terms
| $1.40 | $4.40 | undeclared | 98.07% |
| Fireworksfireworks Endpoint terms
| $1.40 | $4.40 | undeclared | 99.44% |
| Cloudflarecloudflare Endpoint terms
| $1.40 | $4.40 | undeclared | 98.29% |
| Mistralmistral/nvfp4 Endpoint terms
| $1.54 | $4.84 | nvfp4 | 99.71% |
| BaseTenbaseten/fast Endpoint terms
| $2.10 | $6.60 | fp8 | 99.04% |
| BaseTenbaseten/fast Endpoint terms
| $2.10 | $6.60 | fp8 | 99.30% |
| Fireworksfireworks/fast Endpoint terms
| $2.10 | $6.60 | undeclared | 99.62% |
| Alibabaalibaba/fast Endpoint terms
| $2.80 | $8.80 | undeclared | 99.40% |
Independent scores
| LMArena Elo | 1471.4max |
|---|---|
| LMArena row | glm-5.3-max |
LMArena Elo under CC BY 4.0. Absence of another board is not a score of zero.
Best rankings by task
- SVG #8 1293
- Python pptxslides #8 1255
- Htmlslides #9 1189
- 3D #14 1332
- Code categories #15 1304
- UI components #15 1314
- Game development #17 1311
- Websites #18 1295
Per-task Elo and rank from Design Arena, via OpenRouter’s model feed. The rank is the feed’s, and it counts models OpenRouter does not list.
Capability
| Context window | 1M |
|---|---|
| Max output | 943K |
| Input modes | text |
| Tool use | yes |
| Reasoning | always on |
| Open weights | yes |
Provenance
| Price source | Published record of 2026-10-06, under confirmation since 2026-10-12 |
|---|---|
| Fetched | 2026-10-21 |
| Held | Price and provider offers kept at the published record of 2026-10-06 until a second day confirms the newer upstream value, or after 3 days until one day’s two readings agree. |
| Quality data | verified |
| Cross-checked | vendor page |
Our take
editorial — not a measurementGLM is a candidate for text-only workloads where selected-workload cost and task quality can be evaluated together. Use the current proof and benchmark evidence to judge measured price and capability. Do not infer downloadable weights or licence rights from the broader GLM family; verify the exact release and endpoint.
Strengths
- Text input, tools and reasoning effort are listed
- Cached-input pricing is recorded
Weaknesses
- Family history does not establish this release’s weight or licence status
- Text-only input is a migration constraint
Reach for it when
- Text-only workload evaluations
- Teams that compare cost per completed task
Avoid it if
- You require multimodal input
- Self-hosting is required but the exact weights and licence are not established
Sources: openrouter.ai
GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...
Questions this page answers
What does GLM 5.3 cost?
$0.200/M in, $3.40/M out: the published record of 2026-10-06, under confirmation since 2026-10-12. That is the API row, not a subscription.
Does GLM 5.3 have an independent quality score?
GLM 5.3 has an LMArena score in this catalogue.
What beats GLM 5.3?
MiMo-V2.6-Pro has an equal-or-higher measured score and an equal-or-lower price, with capability trade-offs.
Does this page use Artificial Analysis scores for GLM 5.3?
No. Artificial Analysis figures are not published here. Quality on this page is LMArena Elo where a score exists; otherwise the row is unrated.
Is the cheapest GLM 5.3 endpoint the same product?
Headline cheapest is a lower precision. Like-for-like at the best declared precision is $0.190/M from AkashML.
MODEL MONUMENT
Save or share this proof
Badge
The verdict as one SVG, for a README or a docs page. It states this model's dominance status at the balanced workload on the LMArena lens, with the date it was computed, and it is rebuilt with the catalogue — so it changes when the verdict changes, including to one you would rather it did not.
Markdown
[](https://undominated.ai/models/z-ai__glm-5.3/)HTML
<a href="https://undominated.ai/models/z-ai__glm-5.3/"><img src="https://undominated.ai/badge/z-ai__glm-5.3.svg" alt="GLM 5.3 — Undominated.ai dominance verdict" height="36"></a>Direct file: https://undominated.ai/badge/z-ai__glm-5.3.svg — a static SVG, written by scripts/build-badges.mjs on every build, so the copy you embed is never
older than the last deploy.
