KV-cache quantization without any fork (recommended, 2026): upstream llama.cpp/Ollama now cover this natively — use -ctk q8_0 -ctv q8_0 (half KV memory, negligible quality loss: perplexity +0.002–0.05) or -ctk q4_0 -ctv q4_0 (quarter memory, ≈7.6% perplexity increase). In Ollama: OLLAMA_KV_CACHE_TYPE=q8_0 with OLLAMA_FLASH_ATTENTION=1. Keep K and V types symmetric to stay on the fast fused Flash-Attention path. Since April 2026, mainline llama.cpp also applies Hadamard rotation to KV activations (PR #21038), which greatly improves low-bit KV quality (opt-out: LLAMA_ATTN_ROT_DISABLE=1).

The RotorQuant/TurboQuant fork flow below is experimental/legacy: the TurboQuant llama.cpp PR was closed without merging (June 2026) and the fork is unmaintained relative to mainline. It is NOT required to use this model.

Leanstral-RotorQuant-MLX-2bit

2-bit MLX weight-quantized Leanstral-2603 with RotorQuant KV-cache quantization for high-throughput Lean 4 formal proof generation on Apple Silicon.

Leanstral is the first open-source AI agent purpose-built for Lean 4 formal proofs -- generating both executable code and machine-checkable mathematical proofs. This variant combines dual compression: 2-bit MLX weight quantization for aggressive model size reduction plus RotorQuant KV-cache quantization, delivering 5.3x faster prefill and 28% faster decode compared to TurboQuant equivalents.

Overview

This repository provides an aggressively compressed configuration with RotorQuant's superior throughput: MLX 2-bit weight quantization minimizes the static memory footprint, while RotorQuant's rotation-aware KV-cache compression delivers faster prefill and decode than TurboQuant.

Spec	Value
Base model	mistralai/Leanstral-2603
Architecture	Mistral MoE (~119B parameters, 7 consolidated shards)
Weight quantization	2-bit (MLX)
KV-cache quantization	RotorQuant
Weight memory	~30 GB
Prefill speedup	5.3x vs TurboQuant
Decode speedup	28% vs TurboQuant
Runtime	MLX (Apple Silicon)
License	Apache 2.0
Use case	Lean 4 formal verification, theorem proving, mathematical proofs

Quickstart

from mlx_lm import load, generate

model, tokenizer = load("majentik/Leanstral-RotorQuant-MLX-2bit")

prompt = "Prove that for all natural numbers n, n + 0 = n in Lean 4:"
response = generate(
    model,
    tokenizer,
    prompt=prompt,
    max_tokens=512,
)
print(response)

What is RotorQuant?

RotorQuant is an advanced KV-cache quantization method that leverages rotation-aware quantization to achieve superior throughput compared to standard KV-cache compression. By exploiting the rotary positional embedding structure, RotorQuant achieves:

• 5.3x faster prefill -- critical for long Lean 4 proof contexts
• 28% faster decode -- faster token-by-token proof generation
• Equivalent memory savings to TurboQuant with better computational efficiency

Note: 2-bit weight quantization is lossy. Expect some degradation in proof quality compared to the 4-bit variant. For critical formal verification work, prefer the 4-bit or full-precision variants.

Memory Estimates

Component	Estimate
Model weights (2-bit)	~30 GB
KV-cache	Reduced via RotorQuant
Recommended hardware	MacBook Pro M2/M3/M4 Max (64 GB+) or Mac Studio

Lean 4 Use Case

Leanstral excels at:

• Formal verification -- generating machine-checkable proofs of mathematical theorems
• Theorem proving -- interactive and automated proof search in Lean 4
• Code generation -- writing verified Lean 4 programs with correctness guarantees
• Proof repair -- fixing incomplete or broken proof scripts

See Also

• mistralai/Leanstral-2603 -- Base model
• majentik/Leanstral-RotorQuant -- Full-precision weights + RotorQuant KV cache
• majentik/Leanstral-RotorQuant-MLX-4bit -- MLX 4-bit + RotorQuant
• majentik/Leanstral-RotorQuant-MLX-1bit -- MLX 1-bit + RotorQuant
• majentik/Leanstral-TurboQuant-MLX-2bit -- MLX 2-bit + TurboQuant
• RotorQuant repository

Quant trade-off (MLX lane)

Bits	Approx size	Use case	Recommendation
2-bit	~31 GB	Aggressive quantization	Very low-RAM Macs
3-bit	~43 GB	Lossy but small	Low-RAM Macs
4-bit	~50 GB	Balanced default	Recommended for most Macs
5-bit	~60 GB	Higher fidelity	Quality-sensitive
6-bit	~71 GB	Approaching FP16 quality	High-fidelity
8-bit	~90 GB	Near-lossless reference	Fidelity-critical work

(Current variant — 2bit — is bolded.)

Variants in this family

(Showing 8 sibling variants under majentik/leanstral-*. The current variant — RotorQuant-MLX-2bit — is bolded.)

Variant	Runtime	Approx size	Use case
RotorQuant	runtime modifier	n/a	KV-cache root (weight-agnostic)
RotorQuant-MLX-2bit	mlx-lm	card-only	Apple Silicon, smallest
RotorQuant-MLX-4bit	mlx-lm	card-only	Apple Silicon balanced
RotorQuant-MLX-8bit	mlx-lm	card-only	Apple Silicon reference
TurboQuant	runtime modifier	n/a	KV-cache root (weight-agnostic)
TurboQuant-MLX-2bit	mlx-lm	card-only	Apple Silicon, smallest
TurboQuant-MLX-4bit	mlx-lm	card-only	Apple Silicon balanced
TurboQuant-MLX-8bit	mlx-lm	card-only	Apple Silicon reference