
GLM-5
Z.ai
- Okno kontekstowe
- 203K 202 752 tokeny
- Maksymalna długość odpowiedzi
- 131K 131 072 tokeny
- Wejście / Wyjście
- tekst → tekst
Ceny API
USD / 1 mln tokenów
| Wejście | Wyjście | Wejście z pamięci podręcznej |
|---|---|---|
|
|
|
Oś czasu
- · Wersja wag · odnośnik źródłowy
Cena wejścia/wyjścia · USD / 1 mln tokenów
| Model | Odnośniki źródłowe | Cena wejścia/wyjścia |
|---|---|---|
glm-5standard | zai-pricing Trasa API jest niedostępna | $1,00 / $3,20 |
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · smg-project/smg · @hello-alexmcc
Fragment
Atria-Dawn, `glm`, beside Atria-Dawn-Preview: the template writes the GLM-5 shape (`[gMASK]<sop>`, a reasoning-effort system turn, `<|system|>` / `<|user|>` / `<|assistant|>` / `<|observation|>` turns), the tools
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · milabench/milabench · @bouthilx
Fragment
- GLM-5: layout check before launch, loss from the last stage, checkpoint timing and checksums, crash cleanup; `benchmarks/megatron/dev.yaml`, `scripts/trillium/glm/`.
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · maximhq/bifrost · @mickgvirtu
Fragment
is not `schemas.Anthropic`, and the served model -- `glm-5`, `Kimi-K2`, ... -- is neither opus-4.8 nor Fable), so Claude Code's per-turn reminders were hoisted every
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · anomalyco/models.dev · @models-dev-automation[bot]
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · ai-dynamo/frontend-crates · @dundysm
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · ai-dynamo/dynamo · @indrajit96
Fragment
for any DSA model (including DeepSeek V3.2 and GLM-5), or matching `flashmla_kv`/`trtllm` kernels on K3.1's sharded DCP path. Kernels SGLang has not resolved at CLI
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · ROCm/ATOM · @whx-sjtu
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · NVIDIA/TensorRT-LLM · @moraxu
Fragment
class is registered for both `DeepseekV3ForCausalLM` and `DeepseekV32ForCausalLM`; GLM-5 (`GlmMoeDsaForCausalLM`) and Mistral Large 3 subclass it, so exact-class matching excludes them. Kimi K2 declares `DeepseekV3ForCausalLM`
- · Odnośniki źródłowe · GitHub search · Z.ai · pulls · NVIDIA/TensorRT-LLM · @kaiyux
Fragment
- `modeling_glm5_next_vision.py`: the GLM-5 input processor takes the prefix tokenization cache (the fourth commit). Before, it tokenized every text prompt in full and ignored `enable_tokenization_cache`.
Dane techniczne
GLM-5 to model AI od Z.ai z otwartymi wagami, wydany 2026-02-12. Okno kontekstowe wynosi 202 752 tokeny. Maksymalna długość odpowiedzi to 131 072 tokeny. Cena za milion tokenów wynosi $1,00 na wejściu i $3,20 na wyjściu.
- Okno kontekstowe
- 202 752
- Maksymalna długość odpowiedzi
- 131 072
- Wejście
- tekst
- Wyjście
- tekst
- Wagi
- Otwarte
- Wejście · USD / 1 mln tokenów
- $1,00
- Wyjście · USD / 1 mln tokenów
- $3,20
- Wejście z pamięci podręcznej · USD / 1 mln tokenów
- $0,20
Pytania
Jakie jest okno kontekstowe GLM-5?
Okno kontekstowe wynosi 202 752 tokeny. Maksymalna długość odpowiedzi to 131 072 tokeny.
Ile kosztuje GLM-5?
Cena za milion tokenów wynosi $1,00 na wejściu i $3,20 na wyjściu. Wejście z pamięci podręcznej kosztuje $0,20 za milion tokenów.
Czy wagi modelu GLM-5 są publicznie dostępne?
Tak. Z.ai publikuje wagi modelu.
Kiedy wydano GLM-5?
2026-02-12.