RecursiveSI/flint
Text Generation • 40B • Updated • 35.5k • 1.35k
Trinhnc
mrtrinhtn
webbrain-one-322981's profile picture
webbrain-one-537609's profile picture
sattar20's profile picture·
AI & ML interests
None yet
Recent Activity
liked a model 8 days ago
RecursiveSI/flint liked a model 15 days ago
ornith-ai/Ornith-1.5-122B-GGUF liked a model 20 days ago
DavidAU/NVIDIA-Nemotron-3.5-Bolt-Turbo-Brilliance-50x-74B-A9B-24GB-GGUFOrganizations
None yet
liked a
model 8 days ago
liked a
model 15 days ago
liked a
model 20 days ago
reacted to aryan-sabi's post with 👍 21 days ago
Post
3,803
tl;dr: new qwen finetune that simultaneously pushes the frontiers of small model capability and reasoning efficiency.
Hey everyone,
We are Sabi (https://sabi.com), a neuro-AI startup building a BCI + AI wearable that can do agentic actions using just your neural intents.
We publish two post-trained versions of Qwen 3.8 27B called Sabi-Noema and Sabi-Noema-MLA. These models push the capability frontier on <30B models with 2x reduction in verbosity, and are exclusively trained on open source environments.
These models have improved capabilities on coding, agentic and tool calling tasks, as well as hallucinating lesser. Sabi-Noema beats GPT 6 Luna-medium on DeepSWE, placing in between Luna 6 medium and high. It beats Luna-high on agentic and tool calling benchmarks like TB2.1, Tau-3 bench, ClawEval, and BFCL v4. Its hallucination rate is considerably lower than Luna, Qwen and other competing models.
These models are much more reasoning efficient. Sabi-Noema generates 50% fewer tokens per response on TB2.1 and 25% fewer tokens on DeepSWE, compared to base Qwen. Sabi-Noema-MLA goes even further by doing a post-hoc MLA conversion of Qwen's GQA attention. This conversion uses only a calibration dataset and no extra training. With BF16 latents Sabi-Noema-MLA needs 3.6x lower bytes per token. The MLA latents can be further compressed to FP8 with the total KV cache size on TB2.1 to 8.4x smaller with similar performance. To put this in perspective, when using Sabi-Noema-MLA at q4_0, on a 24GB GPU like 3090/4090 you can get a context size of ~534K where base Qwen at FP8 cache would only get to ~150K context.
We are excited to have the open source community play around with these models. We will be releasing more post-trains in the future focused on making <30B models more powerful.
HuggingFace:
Sabi-Noema: sabicap/sabi-noema
Sabi-Noema-MLA: sabicap/sabi-noema-mla
Hey everyone,
We are Sabi (https://sabi.com), a neuro-AI startup building a BCI + AI wearable that can do agentic actions using just your neural intents.
We publish two post-trained versions of Qwen 3.8 27B called Sabi-Noema and Sabi-Noema-MLA. These models push the capability frontier on <30B models with 2x reduction in verbosity, and are exclusively trained on open source environments.
These models have improved capabilities on coding, agentic and tool calling tasks, as well as hallucinating lesser. Sabi-Noema beats GPT 6 Luna-medium on DeepSWE, placing in between Luna 6 medium and high. It beats Luna-high on agentic and tool calling benchmarks like TB2.1, Tau-3 bench, ClawEval, and BFCL v4. Its hallucination rate is considerably lower than Luna, Qwen and other competing models.
These models are much more reasoning efficient. Sabi-Noema generates 50% fewer tokens per response on TB2.1 and 25% fewer tokens on DeepSWE, compared to base Qwen. Sabi-Noema-MLA goes even further by doing a post-hoc MLA conversion of Qwen's GQA attention. This conversion uses only a calibration dataset and no extra training. With BF16 latents Sabi-Noema-MLA needs 3.6x lower bytes per token. The MLA latents can be further compressed to FP8 with the total KV cache size on TB2.1 to 8.4x smaller with similar performance. To put this in perspective, when using Sabi-Noema-MLA at q4_0, on a 24GB GPU like 3090/4090 you can get a context size of ~534K where base Qwen at FP8 cache would only get to ~150K context.
We are excited to have the open source community play around with these models. We will be releasing more post-trains in the future focused on making <30B models more powerful.
HuggingFace:
Sabi-Noema: sabicap/sabi-noema
Sabi-Noema-MLA: sabicap/sabi-noema-mla
- 7 replies
reacted to DavidAU's post with ❤️ 21 days ago
Post
16,095
TURBO-BRILLIANCE 50X - Qwen 3.8 27B, LFM, and more coming soon...
15 AI CONSTRUCTS:
Each a precision engineered "consciousness" to help you in every task, using fewer tokens and generating better answers.
This model has significant internal upgrades under user control that are "on the fly" (you set in your prompt(s) and/or API).
Please read this model card carefully to get the most from it - including all the reasoning and instruct modes (27+27+9 xhigh) that feature model suggestions ("context help") for all use cases.
The Turbo Brilliance system will begin appearing in other model sizes/types shortly (IE: Qwen, Gemma, Mistral, GLM, Llama, ... including our strongest fine tunes too). The "rollout" is now in progress...
DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF
Qwen 3.5 9B The Defiant (1000+likes, 2m+ downloads) [ Limited 12x]:
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF
LFMs [Limited, 12x]:
DavidAU/LFM2.5-2.6B-Qwen3.8-Turbo-Brilliance-Power-X12-NEO-MAX-GGUF
DavidAU/LFM2.5-8B-A1B-Qwen3.8-Turbo-Brilliance-Power-X12-NEO-MAX-GGUF
PS: Qwen Next, LFM2.5 24B-A2B, Qwen 3.6 35B-A3B and others in testing now.
COLLECTION:
https://huggingface.co/collections/DavidAU/turbo-brilliance-ai-constructs-50x-reasoning-instruct
Please see our support page for breakdown of models (sort), info on projects and a lot more information:
DavidAU/support
15 AI CONSTRUCTS:
Each a precision engineered "consciousness" to help you in every task, using fewer tokens and generating better answers.
This model has significant internal upgrades under user control that are "on the fly" (you set in your prompt(s) and/or API).
Please read this model card carefully to get the most from it - including all the reasoning and instruct modes (27+27+9 xhigh) that feature model suggestions ("context help") for all use cases.
The Turbo Brilliance system will begin appearing in other model sizes/types shortly (IE: Qwen, Gemma, Mistral, GLM, Llama, ... including our strongest fine tunes too). The "rollout" is now in progress...
DavidAU/Qwen3.8-27B-Turbo-Brilliance-50X-Reasoning-Instruct-NEO-MAX-MTP-GGUF
Qwen 3.5 9B The Defiant (1000+likes, 2m+ downloads) [ Limited 12x]:
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF
LFMs [Limited, 12x]:
DavidAU/LFM2.5-2.6B-Qwen3.8-Turbo-Brilliance-Power-X12-NEO-MAX-GGUF
DavidAU/LFM2.5-8B-A1B-Qwen3.8-Turbo-Brilliance-Power-X12-NEO-MAX-GGUF
PS: Qwen Next, LFM2.5 24B-A2B, Qwen 3.6 35B-A3B and others in testing now.
COLLECTION:
https://huggingface.co/collections/DavidAU/turbo-brilliance-ai-constructs-50x-reasoning-instruct
Please see our support page for breakdown of models (sort), info on projects and a lot more information:
DavidAU/support
- 3 replies
reacted to DavidAU's post with 🔥 21 days ago
Post
24,864
Qwen 3.5 9B - The Defiant, 27B power ; now with Qwen 3.8 Reasoning modes.
640 ARC-C for both 8bit and 4bit. Model exceeds 7 of 7 benchmarks for Qwen 3.5 9B, Qwen3.5 27B, Qwen3.6 35B-A3B, and meets Qwen 3.6 27B in some cases... and it does so in 4bit and 8bit. Regular and MTP (fast) NEO IMATRIX GGUFs provided. (this model is part of the Qwen 3.6 27B Fable Fusion 711 pipelines: 2200+ likes, 3 million + downloads)
NEW - Qwen 3.8 Reasoning Modes: 2 MTP quants (Q6/Q8) Now with 5 reasoning modes (2 new - Spoon / Einstein), and 5 instruct modes (2 new - Spoon / Einstein, all use ZERO REASONING TOKENS) all switchable on the fly via API, direct and "in chat" (yes - model control at the chat/message level). Model name has "plusIQ" in the name.
(there is also a extra robust "tools" version too.)
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF
PS: I have posted a 22k output example using "spoon" mode too.
UPDATE:
This model (and a few more) will soon have 12 reasoning and 12 instruct modes plus interactive help, model embedded system to select the best reasoning/instruct mode[s] for all use cases.
Final testing is in progress...
640 ARC-C for both 8bit and 4bit. Model exceeds 7 of 7 benchmarks for Qwen 3.5 9B, Qwen3.5 27B, Qwen3.6 35B-A3B, and meets Qwen 3.6 27B in some cases... and it does so in 4bit and 8bit. Regular and MTP (fast) NEO IMATRIX GGUFs provided. (this model is part of the Qwen 3.6 27B Fable Fusion 711 pipelines: 2200+ likes, 3 million + downloads)
NEW - Qwen 3.8 Reasoning Modes: 2 MTP quants (Q6/Q8) Now with 5 reasoning modes (2 new - Spoon / Einstein), and 5 instruct modes (2 new - Spoon / Einstein, all use ZERO REASONING TOKENS) all switchable on the fly via API, direct and "in chat" (yes - model control at the chat/message level). Model name has "plusIQ" in the name.
(there is also a extra robust "tools" version too.)
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF
PS: I have posted a 22k output example using "spoon" mode too.
UPDATE:
This model (and a few more) will soon have 12 reasoning and 12 instruct modes plus interactive help, model embedded system to select the best reasoning/instruct mode[s] for all use cases.
Final testing is in progress...
- 6 replies
liked a
model 21 days ago
reacted to yanghaojin's post with 🔥 21 days ago
Post
3,017
Dear all,
we are happy to share that gbx v0.8.0, our local inference engine for Apple Silicon, runs the 105 GB Qwen3.8-Flash-Next on a 48 GB MacBook Pro by streaming MoE experts from the SSD. When the model fits in memory, it decodes up to 50% faster than omlx, Ollama and mlx-lm on the same Mac, with output identical token for token.
Full write-up: https://medium.com/@semamediadata/running-a-105-gb-qwen-on-a-48-gb-macbook-pro-f116fe493a49
Download gbx: https://github.com/GreenBitAI/gbx-lm/releases
Models: https://huggingface.co/collections/GreenBitAI/greenbitai-for-apple-silicon
we are happy to share that gbx v0.8.0, our local inference engine for Apple Silicon, runs the 105 GB Qwen3.8-Flash-Next on a 48 GB MacBook Pro by streaming MoE experts from the SSD. When the model fits in memory, it decodes up to 50% faster than omlx, Ollama and mlx-lm on the same Mac, with output identical token for token.
Full write-up: https://medium.com/@semamediadata/running-a-105-gb-qwen-on-a-48-gb-macbook-pro-f116fe493a49
Download gbx: https://github.com/GreenBitAI/gbx-lm/releases
Models: https://huggingface.co/collections/GreenBitAI/greenbitai-for-apple-silicon
- 2 replies
New activity in DavidAU/Qwen3.8-Flash-Next-Turbo-Brilliance-50x-Reasoning-Instruct-GSQ-RCO-GGUF 21 days ago
need more ver gsq for coder moe pls boss
2
#1 opened 21 days ago
by mrtrinhtn
liked a
model 21 days ago
liked 2
models 22 days ago
liked a
model 23 days ago
liked a
model 29 days ago
liked 5
models about 1 month ago
pfeifferj/Ornith-1.5-35B-A3B-GSQ-RCO-GGUF
Text Generation • 35B • Updated • 12.6k • 25
pfeifferj/DeepSeek-V4.1-Flash-GSQ-RCO-GGUF
Image-Text-to-Text • 748B • Updated • 22.7k • 22
pfeifferj/GLM-5.3-Flash-GSQ-RCO-GGUF
Image-Text-to-Text • 313B • Updated • 20.9k • 36
pfeifferj/Qwen3.8-Flash-Next-GSQ-RCO-GGUF
Image-Text-to-Text • 125B • Updated • 40.4k • 53
DavidAU/Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic-NEO-IMATRIX-MAX-MTP-GGUF
Image-Text-to-Text • 9B • Updated • 2.01M • 1.41k
liked a
model 3 months ago