Skip to content

Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.

C++
NameLast commit date
..
adapt_async_default.hpp
arch_defaults.hpp
auto_cache.hpp
batch_rows.hpp
batch_telemetry.hpp
conversation_buffer.hpp
conversation_cache.hpp
conversation_file.hpp
conversation_memory.hpp
conversation_snapshot.hpp
coupled_draft.hpp
defaults_026.hpp
device.hpp
dflash2.hpp
dma_batch.hpp
draft_cap.hpp
emulate.hpp
exchange_storage.hpp
expert_cache.hpp
expert_nibble.hpp
expert_source.hpp
family_layout.hpp
foresight_swap.hpp
gated_proposal.hpp
geometry.hpp
gfx12_platform_defaults.hpp
graph.hpp
hit_hook.hpp
host_graph_launch.hpp
layer.hpp
layout.hpp
logit_bias.hpp
mtp.hpp
native_dense.hpp
native_head.hpp
nv_defaults.hpp
on_device.hpp
peer_experts.hpp
pin_budget.hpp
pinned_ring.hpp
pinned.hpp
ple_residency.hpp
progress.hpp
remote_expert_opt.hpp
remote_experts.hpp
session_save_reclaim.hpp
session.hpp
sm120_series.hpp
sm75_series.hpp
spec_prob.hpp
spec_proposal.hpp
tc_sm75.hpp
uma_cache.hpp
verify.hpp
vmm.hpp
weights.hpp
window_evict.hpp