BBuf/vllm-mimo-v2-flash-optimization
skills/model-optimization/vllm/vllm-mimo-v2-flash-optimization/SKILL.md
PR-backed optimization manual for MiMo-V2 / MiMo-V2-Flash / MiMo-V2.5 in vLLM. Use when an engineer needs to audit, debug, extend, or document MiMo-V2 inference-centric MoE runtime, MTP behavior, MiMo-V2.5 Pro/Omni support, and the transition from older MiMo checkpoints in vLLM.
$ install --global
skillsauthnpx skillsauth add BBuf/AI-Infra-Auto-Driven-SKILLS vllm-mimo-v2-flash-optimizationInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 2, 2026, 2:39 AM55.7s2 files scanned
SKILL.md
- name:
- vllm-mimo-v2-flash-optimization
- description:
- PR-backed optimization manual for MiMo-V2 / MiMo-V2-Flash / MiMo-V2.5 in vLLM. Use when an engineer needs to audit, debug, extend, or document MiMo-V2 inference-centric MoE runtime, MTP behavior, MiMo-V2.5 Pro/Omni support, and the transition from older MiMo checkpoints in vLLM.
vLLM MiMo-V2 / MiMo-V2.5 Optimization
Overview
This skill covers the MiMo-V2 family in vLLM: historical MiMo-V2-Flash,
current mimo_v2.py runtime, MTP behavior, MiMo-V2.5-Pro, MiMo-V2.5-Omni, and
the transition from older MiMo checkpoints.
Evidence snapshot:
- vLLM
origin/main:fd74c90d9on2026-04-27 - Latest family update:
#40967added MiMo-V2.5 Pro and Omni support, moving the active runtime surface tomimo_v2.py,mimo_v2_mtp.py, andmimo_v2_omni.pywhile retaining the historical Flash architecture alias. - Canonical PR notes:
references/pr-history.md - History mirrors:
model-pr-optimization-history/vllm/mimo-v2-flash/README.zh.mdandREADME.en.md
Non-Negotiable Evidence Rule
Use skills/model-optimization/model-pr-diff-dossier/SKILL.md as the production bar.
Every PR cited for this family must be based on diff reading, not only PR titles.
Runtime Surfaces
- Current runtime:
vllm/vllm/model_executor/models/mimo_v2.py - Current MTP runtime:
vllm/vllm/model_executor/models/mimo_v2_mtp.py - Current Omni runtime:
vllm/vllm/model_executor/models/mimo_v2_omni.py - Historical Flash runtime:
vllm/vllm/model_executor/models/mimo_v2_flash.py vllm/vllm/model_executor/models/mimo.pyvllm/vllm/model_executor/models/mimo_mtp.py
Current Main Summary
- MiMo-V2-Flash is the historical throughput-oriented MoE landing in vLLM.
- Current mainline maps
MiMoV2FlashForCausalLMandMiMoV2ProForCausalLMtomimo_v2, with MiMo-V2.5 Omni handled throughmimo_v2_omni. - MTP correctness, registry aliases, multimodal processor coverage, and the split between older MiMo checkpoints and V2.5 checkpoints are the key maintenance points.
Key Landed PRs
- #17433
Support MiMo-7B inference with MTP: Historical base for the MiMo family. - #25136
Fix MTP inference path for MiMo-7B model: Closed a concrete draft-path bug. - #30836
Add MiMo-V2-Flash support: Landed the dedicated V2-Flash runtime. - #40045
use diff kv backend for mimo v2 flash: Added the DiffKV attention backend path. - #40967
Add MiMo-V2.5 support: Added Pro, Omni, MTP, registry, and tests/docs coverage for the current family.
Validation Lanes
- Startup on current MiMo-V2-Flash, MiMo-V2.5-Pro, and MiMo-V2.5-Omni checkpoints where access is available.
- Re-run the parser, quantization, MTP, or multimodal lane that matches this family.
- Re-check tests after touching loader or processor code.
References
references/pr-history.md: diff-reviewed MiMo-V2 family PR cards; it includes historical Flash-file cards and current MiMo-V2.5 updates.
Related Skills
BBuf/sglang-humanize-review
development
VerifiedTrustedCommunity
Perform SGLang code review in the style of human maintainers by consulting the full non-agent PR review episode corpus from project start through the latest refresh (June 2026), including inline review threads, top-level PR comments, review submissions, original multilingual text, and multi-round discussions. Use when reviewing SGLang PRs, diffs, patches, or local changes for correctness, tests, performance, GPU/runtime risks, API compatibility, and maintainability.
531SKILL.mdUpdated May 21, 2026
BBuf/model-pr-history-knowledge
documentation
VerifiedTrustedCommunity
Use when an SGLang, vLLM, or TensorRT-LLM serving/model optimization task needs prior model-family PR evidence. Query and read the PR-driven history docs under model-pr-optimization-history before choosing source paths, fast paths, kernel/fusion ideas, regression risks, or validation lanes.
531SKILL.mdUpdated May 17, 2026
BBuf/vllm-sota-humanize-loop
development
VerifiedTrustedCommunity
Run an autonomous Humanize-governed vLLM SOTA performance loop for one LLM model: first perform the fixed fair vLLM/SGLang/TensorRT-LLM deployment search and benchmark, then start one RLCR loop that repeatedly decides the gap, profiles the current bottleneck, runs layer/kernel pipeline analysis, patches vLLM code, optionally uses ncu-report-skill for kernel evidence, and revalidates until vLLM matches or beats the best observed framework under the same workload and SLA.
423SKILL.mdUpdated May 27, 2026
BBuf/llm-pipeline-analysis
devops
VerifiedTrustedCommunity
Inspect LLM torch profiler traces at forward-pass, layer, and kernel level. Use when you need layer timings, anchor-kernel boundaries, representative kernel flows, or Perfetto time ranges.
423SKILL.mdUpdated May 21, 2026