Issues

• Error: 400: {"type":"invalid_request_error","message":"prompt (38963 tokens) + max tokens (226281) exceeds the context (262144); requests are never truncated"}

  Niko1221/Strata#545 · Filter by author avarakin opened on Oct 2, 2026

• Please run several security audits

  Niko1221/Strata#544 · Filter by author bennmann opened on Oct 2, 2026

• Can we add OpenCode config.jsonc to docs?

  Niko1221/Strata#543 · Filter by author frederikhors opened on Oct 2, 2026

• sm_120 + IQ packs: ~4x prefill regression 0.1.34-0.1.36 when libcudart's ABI is mismatched - the #420 fits() gate silently disables MMQ (smpbo reads 1)

  Niko1221/Strata#542 · Filter by author gravitomagnetic opened on Oct 2, 2026
  1

• AMD/HIP (R9700, engine 0.1.36): batched prompt read stalls at a repeatable token; STRATA_PLE_BATCH=0 avoids it

  Niko1221/Strata#541 · Filter by author icodebot opened on Oct 2, 2026

• Strata support AMD's RDNA2, specifically the RX 6900 XT.

  Niko1221/Strata#539 · Filter by author aidoluiz opened on Oct 2, 2026

• Literal </think> quoted in reasoning yields empty stop or leaked reasoning (v0.1.37)

  Niko1221/Strata#537 · Filter by author fenrir-labs76 opened on Oct 2, 2026
  2

• Thinking speed suddenly drops to 0.5 token/s

  Niko1221/Strata#535 · Filter by author sinand99 opened on Oct 2, 2026

• Performance, RTX3060 vs RX7800XT

  Niko1221/Strata#534 · Filter by author blastbeng opened on Oct 2, 2026
  1

• Hot VRAM resize: let the engine yield VRAM to other apps and take it back

  Niko1221/Strata#533 · Filter by author AXKore opened on Oct 2, 2026

• high/xhigh reasoning_effort can silently run to max_tokens with empty content when reasoning_budget_tokens isn't set

  Niko1221/Strata#530 · Filter by author maxlippe-leonardo opened on Oct 2, 2026

• Conversation cache: decode drops ~4-6x on continued conversations (0.1.36, Windows, RTX 5090, IQ3_XXS)

  Niko1221/Strata#528 · Filter by author NikolaFC opened on Oct 2, 2026

• Hello

  Niko1221/Strata#522 · Filter by author ebdellifakher018-rgb opened on Oct 2, 2026

• [2x RTX PRO 6000] Would this approach allow me to run full bf16 model?

  Niko1221/Strata#520 · Filter by author lulinxuan opened on Oct 2, 2026

• 0.1.36 on an RTX 5090: STRATA_PF_FUSED=1 speeds up IQ3_XXS prompts 13-23%, and a ~550-token prompt spends ~470 ms in the batched prompt path

  Niko1221/Strata#519 · Filter by author brenoperucchi opened on Oct 2, 2026
  2

• Feature request: MacOS M-Chip support

  Niko1221/Strata#518 · Filter by author xyd945 opened on Oct 2, 2026

• new error raise RuntimeError("the engine exited before it was ready" + (f" (see {log})" if log else "") +

  Niko1221/Strata#517 · Filter by author XeonG opened on Oct 2, 2026
  4

• desktop (KDE/Wayland): compositor VRAM exhaustion when expert-cache auto fills a 24 GB card - pin framebuffer -12, KWin graphics reset

  Niko1221/Strata#516 · Filter by author xyzzing opened on Oct 2, 2026

• Feature request: Intel Arc B390 / Panther Lake iGPU support on Windows (64 GB shared RAM)

  Niko1221/Strata#515 · Filter by author JerroldH opened on Oct 2, 2026

• Feature Request: use AMD iGPU instead of CPU for offloading

  Niko1221/Strata#514 · Filter by author Vitaliy86 opened on Oct 2, 2026

• Feature Request: experimental support for Qwen3.8-Whittle-MoE-27B-A17.8B

  Niko1221/Strata#513 · Filter by author Vitaliy86 opened on Oct 2, 2026

• Engine stopped unexpectedly

  Niko1221/Strata#511 · Filter by author sinand99 opened on Oct 2, 2026

• RuntimeError("the engine exited before it was ready" + (f" (see {log})" if log else ""))

  Niko1221/Strata#496 · Filter by author AlexBekhtin opened on Oct 2, 2026
  7

• Heterogeneous multi-GPU roles: whole model on one GPU, the MTP draft head on the other (+ test harness)

  Niko1221/Strata#490 · Filter by author hireymage opened on Oct 2, 2026
  3

• [2x RTX 2080 Ti ] 262K context can OOM the weight arena / hard-freeze host; 204800 is stable

  Niko1221/Strata#486 · Filter by author wxaef opened on Oct 2, 2026
  3