Run Recursive Flint on M6

The LLMCheck index estimates Recursive Flint at ~61 tok/s on an M6 at Q4_K_M, scaled from its community-measured M3 Pro run with the published bandwidth model, not a measured run on this chip. It needs about 20 GB of unified memory. It is Recursive's 39.8B mixture-of-experts model (2.5B active), released under the Apache 2.0 licence.

Speed
estimated~61
tok/s
First Token
—
seconds
RAM Needed
20
GB minimum
Engine
—
estimate
Want to run Recursive Flint faster — or step up to bigger models? Find your Mac →

The figure

This is an LLMCheck index estimate from its published two-term bandwidth model, not a measured run; the index labels it so everywhere it appears. The bandwidth formula’s typical error for small MoEs (≤40B total) is ±15%, against 88 measured runs of 15 models (how it is computed). Own this Mac? Submit a real run →

MetricValue
Tokens per secondestimated~61 tok/s
Provenanceestimate · how estimates are made
QuantizationQ4_K_M
Minimum RAM20 GB
Parameters39.8B MoE (2.5B active)
Chip bandwidth170 GB/s

Every figure for Recursive Flint on the M6

Each row as the index holds it, with its source. “Mac RAM” is the memory of the Mac a run used, not the model’s requirement (20 GB). The headline figure follows the reference rule: the lower-median measured plain-decode run at 5 bits or fewer (a run whose quant was not stated counts only when no stated-quant run qualifies), else the estimate; a run marked “not used to anchor estimates” is listed here but never becomes the headline.

SpeedQuantEngineMac RAMContextSourceNote
community83.5 tok/smixed 3/4-bitMLX32 GB1komlx.ai12-core GPU. oMLX v0.7.3, MTP/DFlash off; flint-mixed-3-4bit, a build with no public repo and no stated bits per weight, so it does not anchor; 17.1 GB peak. 80.8 tok/s at 4k.; not used to anchor estimates
Provenance key
sourced
Published figure — vendor or press review
estimated
community
Linked individual run — blog, forum or submission

Speed by context depth: Recursive Flint on the M6

Measured figures only, each from its source’s own session; the index does not estimate long-context or prefill speed, so a missing cell is —. Each @ column is plain decode with that many tokens already in the prompt, from the same session as the row’s short-context figure: decode slows as the KV cache it reads grows. Prefill is prompt processing, the tokens per second a runtime reads a prompt at; with long agent contexts it sets how long you wait before the first token.

RunDecode (row)@4K@8KPrefillTTFT
MLX · mixed 3/4-bit · 32 GB Mac · omlx.aicommunity83.5 tok/s
1k
community80.8 tok/scommunity75.4 tok/scommunity1313 tok/s
at 8K
community1.249 s

Setup: run Recursive Flint on a Mac

Ollama — install it (brew install ollama or the app from ollama.com), then pull the model by its tag, checked on ollama.com on 2026-11-14:

ollama run flint:40b

MLX — pick a build from Hugging Face’s MLX listings, then run it with its repo id:

pip install -U mlx-lm
mlx_lm.generate --model <repo-id> --prompt "Hello"

Downloads: Ollama page → · Hugging Face model card →

Prefer a GUI? LM Studio downloads GGUF and MLX builds in one click. Step by step: install Ollama on a Mac.

Recursive Flint on other Apple Silicon chips

ChipSpeedQuantEngineBandwidth
M5 Maxcommunity123 tok/s4-bitMLX614 GB/s
M5 Procommunity92 tok/sQ4_1llama.cpp307 GB/s
M3 Ultracommunity91 tok/s4-bitMLX819 GB/s
M4 Procommunity80 tok/sQ4_1llama.cpp273 GB/s
M3 Procommunity51 tok/sQ4_1llama.cpp150 GB/s

About Recursive Flint

Recursive's Apache 2.0 coding-agent MoE (21 Oct 2026): 39.8B total, ~2.5B active, 262K context, text only, always reasons. Its card reports SWE-bench Verified 74.2 and Terminal-Bench 2.1 54.9 (its SWE-bench Pro v2 89.7 is an internal mini-swe-agent run, not comparable to SEAL's); the index's capability figure is estimated from these vendor coding benchmarks only. Stock llama.cpp b11803+, Ollama (flint:40b) and LM Studio run it, though llama.cpp's expert-scaling workaround costs 10–30% for now and issue #31000 (NVFP4 tool calls) is open. A ~18 GB 3-bit build fits a 24 GB Mac; community llama.cpp runs measured 129.6 tok/s on a full 40-core M5 Max (b11871), 92.4 on an M5 Pro and 79.6 on an M4 Pro (Q4_1).

Apache 2.0 licence · 39.8B MoE · 262K context · LLMCheck capability 37/50 · see its leaderboard rank →

Memory by context length

Estimated, not measured: the index’s minimum for Recursive Flint (20 GB — the weights at the quant that fits, plus ~2 GB) plus the KV cache for that many tokens, computed from its config.json. 14 full layers + 35 sliding-window layers (window 513). A 16-bit KV cache is the default in llama.cpp, Ollama, LM Studio and mlx-lm; a q8 cache halves the KV part. Coding agents hold 64K tokens or more. Runtimes also add a buffer that grows with context, so leave a few GB spare. macOS lets the GPU use only part of a Mac’s memory by default (figures for an M6 Mac, which runs macOS 26 or later): “fits” means within the GPU memory a Mac of that tier can use with the limit raised, a tier tagged raise GPU limit fits only after sudo sysctl iogpu.wired_limit_mb raises it (the tooltip has the exact command), and the last column is the smallest Mac memory size whose default limit holds it (the GPU memory limit).

ContextKV cache (16-bit)RAM neededRAM, q8 KVSmallest tier that fitsWithin the default GPU limit
8K~0.3 GB~20.1 GB~20.1 GB32 GB Macsame
32K~1.0 GB~20.8 GB~20.4 GB32 GB Macsame
64K (coding-agent context)~1.9 GB~21.8 GB~20.9 GB32 GB Macsame
128K~3.8 GB~23.6 GB~21.8 GB32 GB Macsame

System requirements

Macs with the M6

🛒 Get a Mac that runs Recursive Flint

Recursive Flint needs about 20 GB of unified memory. These Apple Silicon Macs have enough to run it. Apple builds the higher-memory configurations to order; Amazon does not stock them.

As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep the LLMCheck index free.

Compare More Models

See how Recursive Flint stacks up against other models on your specific Mac hardware.

Open Compare Tool Full Leaderboard