family member
GLM-4.7 logo

GLM-4.7 FLASH

45 tokens a second on Linux · 2× Tesla T4 15 GB · 4 CPUs · 31 GB RAM. About $5.62 per million tokens here, against OpenRouter's $0.40: the API is 14× cheaper. Run it here for privacy or to keep work on your own machines, not to save money.

1 build of GLM-4.7 FLASH, 14 signed runs on real machines: different packagers, precisions and engines, each a different file with its own numbers.

213tok/s, fastest machine
$5.62per M tokens · $0.40 on an API
128Kcontext proved by recall
31.1 GBmemory, smallest build
14providers measured it
Run it on your gram Install a gram with one command. It works through the catalog on your own machine and signs every result. 14 providers have measured GLM-4.7 FLASH so far.Use it in your own room Ask Evo in your gramx and it runs on a machine that proved it, at the metered $5.62 per million tokens. On your own machines it costs you nothing.Compare every build Put it beside every other model in one table: speed, price, context and the API price, one row each.
best on ollama · 1 build

trained 198K · proved 128K context, by unsloth Q4_K_M

213.3 tok/s peak, by unsloth Q4_K_M

Speed against price, every machine

Up and left is faster and cheaper · the line joins machines no other beats on both · price on a log scale · dashed: OpenRouter

$1$10USD per million tokenstok/sLinux · 2× Tesla T4 15 GB · 4 CPUs · 31 GB RAM · 4-bit by unsloth · 44.98 tok/s · $5.6235/MLinux · L4 23 GB · 12 CPUs · 53 GB RAM · 4-bit by unsloth · 40.58 tok/s · $5.6652/MLinux · no GPU · 4 CPUs · 31 GB RAM · 4-bit by unsloth · 3.55 tok/s · $12.5701/MLinux · no GPU · 4 CPUs · 31 GB RAM · 4-bit by unsloth · 3.21 tok/s · $13.9098/M

Same build, different machines

Q4_K_M by unsloth on 14 machines: the fastest decodes 169× as fast as the slowest. One dot per machine, tokens per second.

  1. NVIDIA RTX PRO 6000 Blackwell Server Edition×1213 tok/s
  2. 2× Tesla T4×259 tok/s
  3. NVIDIA L4×151 tok/s
  4. Linux, CPU only · 4 CPUs×103.6 tok/s

What it fits on

The smallest build needs about 31.1 GB; 2 of 18 kinds of machine on the network can hold it. Bars: each kind's memory (its card's, else RAM) · ticks: every build, at the size it held plus a fifth for the prompt.

  1. Linux, CPU only · 2 CPUs1 GB RAMnone
  2. NVIDIA GeForce RTX 3050 Laptop GPU4 GB on the cardnone
  3. NVIDIA GeForce GTX 16504 GB on the cardnone
  4. AMD Radeon(TM) Vega 8 Mobile Graphics7 GB RAMnone
  5. Windows, CPU only · 12 CPUs7 GB RAMnone
  6. Apple silicon8 GB RAMnone
  7. Apple M18 GB RAMnone
  8. Mac, CPU only · 8 CPUs8 GB RAMnone
  9. NVIDIA GeForce RTX 3080 Ti12 GB on the cardnone
  10. Tesla T415 GB on the cardnone
  11. Intel(R) UHD Graphics 73016 GB RAMnone
  12. Intel(R) Iris(R) Xe Graphics16 GB RAMnone
  13. Apple M416 GB RAMnone
  14. Apple M116 GB RAMnone
  15. NVIDIA L422 GB on the cardnone
  16. 2× Tesla T429 GB on the cardnone
  17. Linux, CPU only · 4 CPUs31 GB RAMevery build
  18. NVIDIA RTX PRO 6000 Blackwell Server Edition95 GB on the cardevery build

How far back it remembers, machine by machine

Filled: the needle came back at that prompt size · outlined: asked and lost · blank: not asked · trained for 198K. 0 walks ended because the model forgot, 5 because the machine could not hold the prompt.

machine · build4K8K16K32K64K128Kwhy it stopped
Gram ·a5be462× Tesla T4 · Q4_K_M by unsloth128K too slow to answer here
Gram ·dae679NVIDIA L4 · Q4_K_M by unsloth128K too slow to answer here
Gram ·662a4eLinux, CPU only · 4 CPUs · Q4_K_M by unsloth4K too slow to answer here
Gram ·e61d3b2× Tesla T4 · Q4_K_M by unsloth4K too slow to answer here
Gram ·e3654fLinux, CPU only · 4 CPUs · Q4_K_M by unsloth4K too slow to answer here

Carbon per million tokens

Lowest build on each machine · 2 of 5 read off a power sensor, the rest modelled from the parts · fewer grams is better

  1. Gram ·dae679342 g2.60 J/tok · sensor
  2. Gram ·a5be46393 g3.68 J/tok · sensor
  3. Gram ·e61d3b655 g11.06 J/tok · modelled
  4. Gram ·e3654f793 g13.41 J/tok · modelled
  5. Gram ·662a4e1.2 kg9.38 J/tok · modelled

Carbon is location-based: signed energy × the yearly average of the grid where the machine signed its place (world grid average (Singapore not in the table) · United States grid average · EU grid average (applied to NL) · world grid average (Australia not in the table); Ember Global Electricity Review 2025 (2024 generation), location-based). Facility overhead is not included. A machine on its own solar is not credited until it can prove it.

Measured by

The people whose machines signed these numbers

and 4 more

proved by at least one build: Tool calling · Tool loop finished · Faithful to tool results · Chained tool calls · Structured output · Sustained output · Extended thinking · Writes code · Reasons about code · Synthesises sources · Follows instructions

Every build

Each its own file with its own page: the machines that ran it and their signatures.

buildenginewhose workmemorycontext provedbest tok/smachines
Q4_K_M by unslothhf.co/unsloth/GLM-4.7-Flash-GGUF:Q4_K_Mollamaquantizer · unsloth31.1 GB128K213.314
derivatives — somebody else's change to these weights

Fine-tunes, merges and abliterations of this member, each its own model with its own page. None of their numbers are counted above.

GLM-4.7-Flash-NVFP4A16not measuredGLM-4.7-Flash-RAM-20GB-MLXnot measuredHuihui-GLM-4.7-Flash-abliterated-mlx-4bitnot measuredGLM-4.7-Flash-24B-A3B-KREAP-Codenot measuredglm-4.7-flashnot measuredGLM-4.7-Flash-Codernot measured