The 30-Day Recap (TL;DR)

A Tie at #1: Why MiMo-V2.6-Flash Is Back

On 24 October MiMo-V2.6-Flash fell to #7 (Score 78). Its leaderboard speed then rested on one plain-decode run on an M5 Ultra, 61.7 tok/s, and the index had set aside a second, 76.4 tok/s, because it came from a 64-core-GPU M5 Ultra with 256 GB, which the index wrongly said Apple does not sell. Both halves of that have changed.

More runs. The oMLX community board holds many more MiMo-V2.6-Flash runs on 256 GB M5 Ultras with multi-token prediction, DFlash and SpecPrefill all off in each page’s own settings. Most were posted between 24 September and 26 October, and the October updates missed them. The index now holds 15 such sessions, eleven on the 80-core GPU and four on the 64-core GPU, from 61.0 to 89.4 tok/s across the RL and MOPD checkpoints at 4-bit and MXFP4, all community-measured. The MOPD checkpoint the catalog lists ran at 78.7 tok/s on oMLX 0.7.0 (run bq9yj6vv).

One row per measurement session. The board lists every run separately, and a single 26 September session posted 21 of them. Counting each run would let one machine outvote everyone else, so the index enters a session once: its 1K-token run, or its shortest context when it has no 1K run, and the session median when that run was repeated. oMLX .dev and rc builds count as released runtimes, as the reference-row rule has said since October.

The 64-core configuration is sold. Apple’s store offers the Mac Studio with a 30-core-CPU, 64-core-GPU M5 Ultra and 256 GB at $9,499 (checked 14 and 15 November). Both GPU versions have the same 1.2 TB/s of memory bandwidth, so under the index’s rules the 64-core part is not a binned, lower-bandwidth configuration, and its four sessions count.

The result. The lower median of the 15 sessions is 76.8 tok/s (a 64-core run on oMLX 0.7.3, 14 November), shown as 77, community-measured. That is worth 19 speed points, so MiMo-V2.6-Flash scores 49 (capability) + 19 (speed) + 3 (memory) + 10 (MIT licence) = 81. Qwen 3.6-35B-A3B scores 38 + 25 + 10 + 8 = 81. Ties break by name, so MiMo-V2.6-Flash is listed first, and every surface that names the #1 says it is a tie. Counting only the 80-core sessions would give 78 tok/s and Score 82, #1 outright; the index has no grounds to drop the 64-core sessions, so it does not.

The figure is physically plausible. With 15B active parameters at about 4.26 bits each, MiMo-V2.6-Flash reads roughly 8 GB per token, which caps an M5 Ultra near 154 tok/s; 77 is half of that, and about 1.3 times the 55–59 tok/s community runs on M3 Ultras.

What it means for buyers: nothing changes below 192 GB. MiMo-V2.6-Flash starts at a 192 GB Mac Studio (its smallest GGUF is 126 GB; 4-bit MLX builds are about 172 GB) and is the pick for 192–512 GB. On a Mac with 32 GB or less the pick is Qwen 3.6-35B-A3B, whose own M5 Max figure moved from 118 tok/s published to 130 community-measured, the lower median of eight runs (118–153).

GLM-4.7-Flash: a Correction, Not a Breakthrough

The month’s biggest move is a fix. Zhipu AI released GLM-4.7-Flash under MIT on 19 January; its model card calls it a 30B-A3B mixture-of-experts model, and its configuration (glm4_moe_lite: 64 routed experts, 4 active per token, plus a shared expert) agrees. The index had listed it as a dense 31B model, so it costed every token as a read of all 31B weights: about 26 tok/s estimated on an M5 Max, Score 61, #61.

A community oMLX run exposed the error: 46.6 tok/s on a 32 GB M6 (oQ4e, multi-token prediction off). A dense 31B model at 4-bit cannot pass about 10 tok/s on the M6’s 170 GB/s. Costed as a MoE with about 3B active, GLM-4.7-Flash estimates at 107 tok/s on an M5 Max, takes the full 25 speed points and scores 80: #3, ahead of Recursive Flint on the alphabetical tie-break. Its 18 GB minimum fits a 24 GB Mac. Two caveats: its M5 Max figure is still an estimate, and its capability score (35) is below Qwen 3.6-35B-A3B’s (38). The rank comes from speed, a small memory footprint and the MIT licence.

Recursive Flint’s First Stock M5 Max Run

Since its 21 October release, Recursive Flint’s M5 Max figure had been an estimate held at about 131 tok/s by a community run on the binned 36 GB M5 Max (460 GB/s), because no qualifying run on a full 40-core M5 Max existed. Now one does: 129.6 tok/s with llama.cpp b11871 and Recursive’s Q4_1 GGUF on a 40-core MacBook Pro 16-inch, from a published benchmark log (community-measured, 26 October). Four oMLX runs on full M5 Max chips joined it (77.4–132.5 tok/s, community), so the leaderboard now shows the lower median of the five: 123 tok/s, community-measured (oMLX, 4-bit). The Score stays 80, because speed points cap at 100 tok/s.

Two kinds of run are shown but not counted: two runs of an unidentified “mixed 3/4-bit” build (163.5 and 147.2 tok/s) with no public repository or stated bits per weight, and custom inference engines in the same log that report roughly twice the llama.cpp figure. The binned 36 GB run (131) now sits above the full chip’s lower median (123); the index shows both, as two real measurements that disagree. On a 24 GB Mac mini M6 Flint is estimated at about 71 tok/s, which makes it the new 24 GB pick.

Open-Source Top 10 (November 2026)

LLMCheck index Top 10 — November 2026.
#ModelScoreLicenseMin RAMSpeed (M5 Max unless noted)
1MiMo-V2.6-Flash81MIT170 GB77 tok/s community (M5 Ultra)
2Qwen 3.6-35B-A3B81Apache 2.024 GB130 tok/s community
3GLM-4.7-Flash80MIT18 GB107 tok/s est.
4Recursive Flint80Apache 2.020 GB123 tok/s community
5KAT-Coder-V2.579Apache 2.024 GB89 tok/s community
6Nemotron 3.5 Lightning79OpenMDW18 GB117 tok/s community
7North Mini Code 1.079Apache 2.021 GB120 tok/s community
8Gemma 4 26B-A4B78Apache 2.018 GB107 tok/s published
9MiniCPM-V 4.7 35B-A3B78Apache 2.023 GB100 tok/s community
10Raon-1 mini preview77Apache 2.018 GB123 tok/s est.

Nine of the ten are mixture-of-experts models of 26–40B total parameters with 2.5–4.3B active, and all nine fit a 24 GB Mac; MiMo-V2.6-Flash is the exception. Eight of the ten speeds are measurements; GLM-4.7-Flash’s and Raon-1 mini preview’s are estimates. New in the Top 10 since 24 October: GLM-4.7-Flash. Out: LFM2.5-24B-A2B (#11, Score 76), still the 64 GB pick. The highest capability score a 24 GB Mac can hold, 46 of 50, is now shared by Qwen3.8-27B (#14, Score 72; 31 tok/s community-measured on an M5 Max) and the new Nanbeige4.3-33B. Every row, with its provenance mark, is on the leaderboard.

The Best Pick for Each RAM Tier

Each tier’s pick is the best-ranked model that fits the tier, reaches 15 tok/s on the tier’s representative Mac and runs on a standard Mac runtime. The 24 and 32 GB tiers are now listed separately.

MemoryPickSpeed on the reference MacScoreNeeds
8 GBMiniCPM5-2B (new; was Nanbeige4.2-3B)~39 tok/s estimated, MacBook Air M1 8 GB693 GB
16 GBGemma 4 12B~17 tok/s estimated, Mac mini M6 16 GB669 GB
24 GBRecursive Flint (was Qwen 3.6-35B-A3B)~71 tok/s estimated, Mac mini M6 24 GB8020 GB
32 GBQwen 3.6-35B-A3B59 tok/s community-measured, Mac mini M6 32 GB8124 GB
64 GBLFM2.5-24B-A2B114 tok/s community-measured, Mac mini M5 Pro 64 GB7617 GB
128 GBQwen 3.6-35B-A3B (was Qwen3.8-Flash-Next)130 tok/s community-measured, Mac Studio M5 Max 128 GB8124 GB
192–256 GBMiMo-V2.6-Flash77 tok/s community-measured, Mac Studio M5 Ultra 256 GB81170 GB
512 GBMiMo-V2.6-Flash77 tok/s community-measured on the M5 Ultra (runs on 256 GB machines)81170 GB

Two picks are close calls on the published suitability formula (capability score plus 0.12 points per tok/s, capped at 15). At 24 GB, Recursive Flint scores 45.5 against Qwen 3.6-35B-A3B’s 45.1; a new 58.7 tok/s community run of Qwen on a 32 GB M6 lowered its M6 figure from 64 to 59 and tipped the balance. At 128 GB, Qwen 3.6-35B-A3B (53.0, its speed bonus capped) edges Qwen3.8-Flash-Next (52.6, 47 tok/s community-measured on an M5 Max). The 128 GB pick is the model that runs best there, not the largest one that fits: the most capable model a 128 GB Mac can hold is GLM-5.3-Flash, at about 27 tok/s estimated. Exact figures for all 54 configurations are on the by-RAM page and the Best-by-Mac pages.

Fourteen New Models, by the Mac They Need

Six of the fourteen were released on or after 24 October. Six more are older releases the October changelog had queued as “seen, not yet evaluated”, and two (K2-Horizon-3.7B and Nemotron 3 Ultra) are catalog gaps the index had missed. Each was checked against its Hugging Face card.

8 GB Macs and up

24–32 GB Macs

48–64 GB Macs

96–128 GB Macs

256 GB Mac Studios

512 GB Mac Studios

Server-class

Considered and Not Added

Fine-tunes and derivatives. A derivative is a checkpoint built on a catalog model: continued training, a supervised or RL fine-tune, a merge, an abliteration or a domain specialisation, or a checkpoint whose architecture is identical to one. The index lists a derivative only when all four of these hold: an accountable organisation publishes it with a licence file; an independent evaluation, or a vendor table against its own base with the harness stated, shows a gain worth at least two capability points; it has a downloadable Mac build on a stock runtime; and it is more than a safety-removal, persona or style change. A lab’s own updated checkpoint replaces its entry instead, as DeepSeek V4.1 Flash 1017 and Ling-3.1-flash-1019 did. None of this month’s derivatives met all four: Arete-27B (a supervised fine-tune of Qwen3.8-27B), Sage-A1.5 (a Qwen 3.6-35B-A3B fine-tune), Atria Dawn (built on GLM-5.2; no licence file, no Mac build of the final weights), Ornith-1.5 (licence file still missing), Hydrogen-1.0 and Nex-N2.6 (GLM-5.3 and GLM-5.3-Flash fine-tunes), Intern-S2-397B (Qwen 3.5 397B-A17B’s configuration with an extended vocabulary) and Ling-3.0-flash-Fin (finance continued training).

No practical Mac path. LongCat-2.5 (1.6T, MIT) is published for NPU deployment and is 1.34 TB even in its smallest format; LongCat-2.0 is superseded and server-class; ERNIE-5.5-Flash-Base is a base checkpoint only, under a licence that excludes the EU, the UK and South Korea; Kimi K3.1 Highspeed is API-only.

Size class and niche. MiniCPM5-1B and Spark-X2.6-1.7B (the index lists their larger siblings); Raon-1 nano (KRAFTON’s fork runtimes only; its upstream llama.cpp pull request, #32200, opened on 13 November); SenseNova-Loop-9B-Preview (CUDA only); and draft, translation, image, OCR and world models.

No weights. Qwen 4: Alibaba’s Qwen/Qwen4-27B repository is not public, and the community “Qwen4-27B” repositories on Hugging Face are empty placeholders. Nothing named GLM-5.5 or GLM-6 is on Hugging Face. Every addition and exclusion is logged in the changelog.

Hardware: M6 MacBook Pro, iMac and a $22,299 Studio

Apple announced three Macs on 27 and 28 October; pre-orders opened on 28 October, and all three shipped on 6 November. None has a new Pro, Max or Ultra chip, and Apple published no generation speeds: its M6 footnotes cite only prompt processing (time to first token up to 1.3 times faster than the M5 MacBook Pro), which says nothing about how fast a model writes.

MacChip and bandwidthMemoryApple priceIndex pick
MacBook Pro 14-inch (M6), newM6; 153 GB/s at 16 GB, 170 at 24–32 GB16, 24, 32 GBfrom $1,999 (1 TB)16 GB: Gemma 4 12B; 24 GB: Recursive Flint; 32 GB: Qwen 3.6-35B-A3B
iMac (M6), newM6, 10- or 12-core GPU; “up to 170 GB/s”16, 24, 32 GBfrom $1,49916, 24 and 32 GB: as the Mac mini M6
MacBook Pro with PROLED XDR, newM5 Pro 307 GB/s; M5 Max 460 or 614 GB/s24–128 GBfrom $2,899 (M5 Pro), $4,499 (M5 Max)as the mini-LED MacBook Pro
MacBook Air (M5), newly listedM5; 153 GB/s16, 24, 32 GBfrom $1,29916 GB: Gemma 4 12B; 24 and 32 GB: Qwen 3.6-35B-A3B
Mac Studio M5 Ultra, 512 GB, now orderable36-core CPU, 80-core GPU only; 1,228 GB/s512 GB$22,299 (1 TB)MiMo-V2.6-Flash

The M6 MacBook Pro has the Mac mini M6’s chip and bandwidth, against 153 GB/s for the M5 MacBook Pro it replaces, so the Mac mini M6 figures carry over: Qwen 3.6-35B-A3B at 59 tok/s community-measured on a 32 GB M6, dense 27B models at about 10. The iMac’s spec page gives only “up to 170GB/s”, with no split by memory; the index assumes 153 GB/s at 16 GB, as Apple specifies for the other M6 Macs. The PROLED MacBook Pro uses the same M5 Pro and M5 Max chips as the mini-LED model, which stays on sale, so it runs models at the same speed; the $400 premium buys the tandem-OLED touch display. The MacBook Air M5 has been on sale since March; the index had skipped it and now has three pages for it.

The 512 GB Mac Studio has been orderable since 28 October, only with the 36-core CPU and 80-core GPU M5 Ultra: $22,299 with 1 TB (checked on Apple’s store on 15 November), $11,500 more than the same chip with 256 GB ($10,799). The 64-core 256 GB model is $9,499. Apple quoted 4–6 business days for delivery on 14 November. What 512 GB buys: 14 more of the index’s models fit than at 256 GB (122 against 108), among them Mistral Large 4, Muse Spark Open, DeepSeek V4.1 Flash, Ling-3.1-flash and Kimi K2.6, but only in 2- to 4-bit-class builds and at about 14–31 tok/s estimated on an M5 Ultra, and no plain-decode run on a 512 GB M5 Ultra had been published by 14 November. M5 Ultra vs M3 Ultra vs M5 Max compares the three Studio chips.

The June price rise. Apple raised Mac prices on 25 June, citing the cost of memory (MacRumors): the 14-inch MacBook Pro with M5 Pro went from $2,199 to $2,499, and the M5 Max model now starts at $4,099. The index’s MacBook prices predated the increase, one by $2,800 (it listed the 128 GB M5 Max MacBook Pro at $3,899; Apple charges $6,699), and several build-to-order Mac Studio and Mac mini prices were approximations. Every current configuration now carries Apple’s list price, checked on 14 November.

Previous generation. The Mac mini M4 and M4 Pro, MacBook Air M4, and the Mac Studio with M4 Max or M3 Ultra are no longer sold new (Apple’s store redirects their configuration pages). The index now labels them previous generation. New Best-by-Mac pages this month: the three M6 MacBook Pros, three M6 iMacs and three M5 MacBook Airs, plus the 48 GB MacBook Pro M5 Pro, the 48 GB Mac Studio M5 Max (at $3,099, the cheapest Mac with 614 GB/s) and the 48 GB Mac mini M4 Pro (used).

Measured This Month

The index added 126 measured rows, 100 community runs and 26 published press or vendor figures, so 319 of its 619 speed rows are now measurements; everything else is the formula’s estimate, labelled as one. The best-documented sources:

Still missing: no review of the M6 MacBook Pro or iMac has published a fully specified generation figure, and nobody has posted a plain-decode run on a 512 GB M5 Ultra.

Runtimes: llama.cpp 1.0, Ollama, MLX

Update, 19 November 2026. Two releases since this edition: oMLX 0.8.0 (18 November) is the current stable release and includes Inkling 2 Small support (pull request #5229), and mlx-vlm 0.8.0 shipped on 16 November. The figures below are as published on 16 November.

What the New Runs Say About the Estimator

For mixture-of-experts models above 40B total parameters, the index multiplies its bandwidth estimate by 0.42, a factor fitted in September to DeepSeek V4 Flash. This month’s large-Qwen runs came in well above it: Qwen 3.5 397B-A17B at 61.6 tok/s against an estimate of 36 (1.7 times), and Qwen3-235B-A22B at 46 against 29 (1.6 times), both on M5 Ultras. Across the 18 large MoEs that now have qualifying runs, measured speed lands at 0.26 to 0.88 of the unscaled formula, median 0.57; the per-model list is on the methodology page.

The index keeps 0.42 anyway, for two reasons. It is meant as a conservative default, and three of the 18 measured below it (DeepSeek V4.1 Flash, Qwen3.8-Flash-Next and Mistral Large 4), so a higher factor would overstate some models. And it now decides little: a model with its own measured rows is scaled from those rows, not from the factor, so 0.42 sets the figure only for large MoEs nobody has measured. Both Qwen models are now anchored on their own runs; Qwen3-235B-A22B’s M5 Max estimate rose from 16 to 26 tok/s. A separate factor for Qwen-family MoEs would need a held-out test before the index adopts one. The small-MoE fit from October and the dense formula are unchanged.

Corrections

What to Watch in December

Nothing here enters the index until weights or measurements exist on a primary source.

Frequently Asked Questions

What is the best open-source local LLM for a Mac in November 2026?

According to the LLMCheck index (14 November 2026), MiMo-V2.6-Flash and Qwen 3.6-35B-A3B tie for #1 at LLMCheck Score 81, and MiMo-V2.6-Flash is listed first on the alphabetical tie-break. MiMo-V2.6-Flash runs at 77 tok/s community-measured on an M5 Ultra but needs a Mac Studio with 192 GB or more. Qwen 3.6-35B-A3B needs only 24 GB and is the best pick on a Mac with 32 GB or less, at 59 tok/s community-measured on a 32 GB Mac mini M6.

What is the best local LLM for a Mac with 16, 24 or 32 GB?

The index's picks are Gemma 4 12B for 16 GB (about 17 tok/s estimated on a 16 GB Mac mini M6), Recursive Flint for 24 GB (about 71 tok/s estimated on a 24 GB Mac mini M6) and Qwen 3.6-35B-A3B for 32 GB (59 tok/s community-measured on a 32 GB Mac mini M6). The 24 GB call is close: Qwen 3.6-35B-A3B also fits and trails Flint by less than half a point on the index's suitability formula.

What is the best local LLM for an 8 GB Mac?

MiniCPM5-2B, new this month: a 2.5B Apache 2.0 model that needs about 3 GB and runs at about 39 tok/s estimated on an 8 GB MacBook Air M1. K2-Horizon-3.7B has a slightly higher capability score but runs about half as fast there, and Ling-3.0-tiny is the fastest option, at about 87 tok/s estimated.

Why is GLM-4.7-Flash suddenly #3?

Because the index fixed an error. GLM-4.7-Flash is a 30B-A3B mixture-of-experts model, according to its Hugging Face card and configuration, but the catalog had listed it as a dense 31B model since January, which cut its speed estimate about fourfold. Costed correctly, it is estimated at 107 tok/s on an M5 Max and scores 80. A community run measured 46.6 tok/s on a 32 GB M6.

Can a Mac run Mistral Large 4?

Only a 512 GB Mac Studio. Its Q2_K build is 381 GB. Hugging Face's launch post reports 15 tok/s generation on a 512 GB M3 Ultra with llama.cpp (published), and the index estimates about 21 tok/s on a 512 GB M5 Ultra. llama.cpp (b11998 and later) and mlx-vlm 0.7.12 support it; stock mlx-lm does not yet. Its new Mistral Ecosystem Open-Weight License requires a separate agreement only for Model-as-a-Service businesses.

Is the 512 GB Mac Studio M5 Ultra worth $22,299?

Only if you need models above 256 GB. Apple has sold it since 28 October, with the 80-core GPU only, for $11,500 more than the 256 GB model with the same chip. It fits 14 more of the index's models than 256 GB, among them Mistral Large 4, Muse Spark Open and DeepSeek V4.1 Flash, but those run at about 14-31 tok/s estimated, and no plain-decode run on a 512 GB M5 Ultra had been published by 14 November. The index's pick for 192-512 GB, MiMo-V2.6-Flash, fits from 192 GB.

Is the M6 MacBook Pro good for local LLMs?

For models that fit in 32 GB, yes, especially small mixture-of-experts models. It has the Mac mini M6's chip, with 153 GB/s at 16 GB and 170 GB/s at 24 or 32 GB, from $1,999. Qwen 3.6-35B-A3B ran at 59 tok/s community-measured on a 32 GB M6, while dense 27B models manage about 10. For larger models, the M5 Pro (307 GB/s) and M5 Max (614 GB/s) MacBook Pros remain the faster choice.