Veizik for Apple Silicon — 2026.10.02-r6.2
Sampling options on `veizik run`. The engines are byte-for-byte the r6.1 engines; only the `veizik` command changed, and sampling uses the sampler the engines already contained.
This page records the state as of 2026-10-09 and is not updated as the catalogue moves. For what runs now, see models.
What this build does
Models in this release
| Model | Task | Context | Status |
|---|---|---|---|
| Qwen2.5-0.5B | Language | 32,768 | Available |
| Qwen2.5-1.5B | Language | 32,768 | Available |
| Qwen2.5-7B | Language | 32,768 | Available |
| Qwen3-1.7B | Language | 32,768 | Available |
| Qwen2.5-0.5B-Instruct | Language | 32,768 | Available |
| Qwen2.5-1.5B-Instruct | Language | 32,768 | Available |
| Qwen2.5-7B-Instruct | Language | 32,768 | Available |
| Qwen3-1.7B-Base | Language | 32,768 | Available |
- Platform
- macOS on Apple Silicon (arm64)
- Install size
- 16.3 MB, 124 files — no dependency to install first
- Verification
- The file name, size and SHA-256 are printed on this page and on the release page, and the two agree. The disk image is signed with a Developer ID and notarised, and carries SHA256SUMS covering its other 49 files, so `shasum -a 256 -c SHA256SUMS` checks the contents after you open it.
- File
- veizik-metal-2026.10.02-r6.2.dmg — no longer downloadable
- Size
- 6,027,105 bytes
- SHA-256
- 959c321ed214fbb7477355235f0c3d50c5d89e0921a509e41f12358b0aac51a0
Verify, install, check
$ shasum -a 256 -c SHA256SUMS $ ./install.sh $ veizik doctor
Full steps are in the documentation.
Fixed issues
- `veizik run` takes `--temperature T` (above 0 turns sampling on), `--top-k K` (default 40 when sampling), `--top-p P` (default 1.0) and `--seed N`. The same seed and settings repeat the same text on the same Mac; without a seed, or with `--seed 0`, every run draws a new one. Values must be plain numbers; anything else is refused before the engine starts.
- Without these options `veizik run` is unchanged: greedy decoding, the same output as r6.1 for the same prompt. `veizik chat` and `veizik bench` stay greedy; `veizik serve` already passed request sampler fields to the engine.
Known issues in 2026.10.02-r6.2
- `/v1/embeddings` answers 501 in this release.
- The serve window is 4,096 tokens per request for every model in this release (prompt plus max_tokens); `veizik run` is not limited this way.
- Qwen3-1.7B returns its reasoning text inside the answer; add `/no_think` to the prompt for a direct reply.
- Offline use: the session key lasts 48 hours after an online start; restarting the Mac ends that window.
- Not supported: Intel Macs, Windows, Linux. Headless activation is not included in this build.