Search the full wire by company, model, lab, or keyword. Every story we have ever aggregated.
Developer reflects on year-long MCP server project's growth; notes improved local model tool-calling capability with Gemma 4 and Qwen 3.6 on consumer hardware.
BeeLlama.cpp fork adds DFlash, TurboQuant, and vision support; runs Qwen 3.6 27B Q5 on RTX 3090 with 200k context at 135 tps.
User reports 1.5–2x speedup running Qwen 27B with MTP optimization on dual AMD MI50 GPUs via llama.cpp.
User achieves 80 tok/sec with 128K context on RTX 4070 Super using Qwen3.6 35B quantization and llama.cpp MTP implementation.
User demonstrates Qwen3.6 27B with Pi coding agent for Archlinux system configuration tasks via natural language.
Community discussion about Qwen model availability and pricing constraints.
Qwen 35B-A3B MoE model runs effectively on RTX 3060 12GB with proper MoE block GPU allocation and 16k-32k context windows.
Developer achieves 80+ tokens/sec on Qwen 3.6-27B with MTP + TurboQuant quantization on single RTX 4090 at 262K context.
Linear probing and activation patching reveal latent planning representations in Qwen3, Gemma-3, Llama-3; future constraints encoded at layer boundaries.
GRPO fails under binary rewards due to gradient starvation when all group responses are correct/wrong; group-mean centering fix demonstrated on Qwen3.5.
Benchmark shows TP=2 pinned to NVLink GPU pairs yields +25–53% throughput vs PCIe on Qwen 3.6 27B; TP=4 degrades performance due to cross-pair PCIe bottleneck.
Qwen3.5/3.6 inference optimization guide: NextN MTP speculative decoding achieves 2.9× speedup on RTX 3090 Ti via llama.cpp with zero quality loss.
Alibaba Qwen releases 35B and 27B model variants; community speculates on 9B and 122B release timing.
Community fine-tune of Qwen 3.6 27B with reduced safety filters released in multiple quantization formats.
Reddit discussion: RTX 5090 vs M5 Max 128GB for local Qwen3.6 27B agentic development—tradeoffs between speed and memory.
User achieves 50 tokens/sec with Qwen 3.6 27B on RTX 3090 using MTP speculative decoding at 100k context.
MTP speculative decoding ported to Qwen 3.6 35B shows modest 2.5-6% speedup vs. 2-2.5x on 27B; architecture may limit gains.
User reports successful local inference with Qwen3.6-35B on AMD R9700 GPU, generating functional code and tests via llama-cpp.
GB10 Solution Atlas, Rust+CUDA inference engine, achieves 100+ tok/s on Qwen 35B; open source, minimal footprint, no Python runtime.
User reports Qwen 3.6 27B in Hermes agent harness successfully handles junior IT tasks, signaling maturity of local model + agent systems.
Comparative study of LoRA and QLoRA fine-tuning on Bashkir, a low-resource Turkic language, using models from DistilGPT2 to Qwen2.5-7B.
User demonstrates Qwen3.6 27B running 200k context on single RTX 5090 with NVFP4 quantization in vLLM, sharing exact configuration and parameters.
Qwen3.6-27B with Multi-Token Prediction achieves 2.5x throughput via Unsloth quantization and llama.cpp integration.
Qwen 3.6 27B achieves 2.5x inference speedup via MTP speculative decoding in llama.cpp; 262k context on 48GB with fixed chat templates.
Empirical quantization degradation analysis for Qwen 3.6 27B across 8 compression levels via chess state-tracking task.
Qwen 27B achieves 54 t/s on V100 GPU with MTP optimization in llama.cpp, nearly 2x baseline speed for code review and tool use tasks.
Community discussion comparing Gemma 4 and Qwen 3.6 model suitability across coding, benchmarks, and agentic workloads.
User reports successful MTP speculative decoding on AMD Strix Halo (AI Max 395) with llama.cpp achieving 60-80 tok/s on Qwen 3.6B GGUF.
Developer benchmarked local Qwen 3.6 27B vs cloud models on 150 real coding tasks, finding local matched cloud 97% on 35% of workload, suggesting cost arbitrage opportunity.