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++
Merge f-028-nerazim: DeepSeek stop ids (test_dsv41 stop-id assertion follows the fix; the arch_from_tokenizer cherry-pick is eng-p1's)
Niko1221d76b50681a7b
diff --cc serve/server.pyindex e05250dd0,7990c47ad..133102dca--- a/serve/server.py+++ b/serve/server.py@@@ -4041,11 -4061,10 +4061,10 @@@ class Serviceself.reasoning_budget_tokens = 0 # #123: the config's default thinking budget (0: none)self.repeat_stop_tokens = REPEAT_STOP_TOKENS # #606: one token this many times in a row ends a reply (0: off)self.anthropic_think_unasked = True # #278: "anthropic_thinking": "on_request" -> False- self.tool_call_recovery = False # opt-in: tool calls in forms next to the template's+ self.tool_call_recovery = True # default on (0.2.8): tool calls in forms next to the template's; config "tool_call_recovery": false turns it offself.images = None # serve/images.py's ImageGen when the config has "images"self.restore_after_load = None # a session file an image job saved before it unloaded- self.stop_ids = set(tokenizer.encode(IM_END, parse_special=True) +- tokenizer.encode("<|endoftext|>", parse_special=True))+ self.stop_ids = end_token_ids(tokenizer)self.family = None # the model family ("deepseek41"): set_family() belowself.parser_cls = OutputParser # reads the model's output (reasoning, content, tool calls)# #567: a prompt re-encodes only what follows the last special token it shares with a recent prompt (the