vLLM v0.12.0 Release Notes Highlights

Highlights

This release features 474 commits from 213 contributors (57 new)！

Breaking Changes: This release includes PyTorch 2.9.0 upgrade (CUDA 12.9), V0 deprecations including xformers backend, and scheduled removals - please review the changelog carefully.

Major Features:

• EAGLE Speculative Decoding Improvements: Multi-step CUDA graph support (#29559), DP>1 support (#26086), and multimodal support with Qwen3VL (#29594).
• Significant Performance Optimizations: 18.1% throughput improvement from batch invariant BMM (#29345), 2.2% throughput improvement from shared experts overlap (#28879).
• AMD ROCm Expansion: DeepSeek v3.2 + SparseMLA support (#26670), FP8 MLA decode (#28032), AITER attention backend (#28701).

Model Support

• New model families: PLaMo-3 (#28834), OpenCUA-7B (#29068), HunyuanOCR (#29327), Mistral Large 3 and Ministral 3 (#29757).
• Format support: Gemma3 GGUF multimodal support (#27772).
• Multimodal enhancements: Qwen3 Omni audio-in-video support (#27721), Eagle3 multimodal support for Qwen3VL (#29594).
• Performance: QwenVL cos/sin cache optimization (#28798).

Engine Core

• GPU Model Runner V2 (Experimental) (#25266): Complete refactoring of model execution pipeline:

  • No "reordering" or complex bookkeeping with persistent batch removal
  • GPU-persistent block tables for better scalability with max_model_len and num_kv_groups
  • Triton-native sampler: no -1 temperature hack, efficient per-request seeds, memory-efficient prompt logprobs
  • Simplified DP and CUDA graph implementations
  • Efficient structured outputs support

• Prefill Context Parallel (PCP) (Preparatory) (#28718): Partitions the sequence dimension during prefill for improved long-sequence inference. Complements existing Decode Context Parallel (DCP). See RFC #25749 for details.

• RLHF Support: Pause and Resume Generation for Asynchronous RL Training (#28037).

• KV Cache Enhancements: Cross-layer KV blocks support (#27743), KV cache residency metrics (#27793).

• Audio support: Audio embeddings support in chat completions (#29059).

• Speculative Decoding:

  • Multi-step Eagle with CUDA graph (#29559)
  • EAGLE DP>1 support (#26086)
  • EAGLE3 heads without use_aux_hidden_states (#27688)
  • Eagle multimodal CUDA graphs with MRoPE (#28896)
  • Logprobs support with spec decode + async scheduling (#29223)

• Configuration: Flexible inputs_embeds_size separate from hidden_size (#29741), --fully-sharded-loras for fused_moe (#28761).

Hardware & Performance

• NVIDIA Performance:

  • Batch invariant BMM optimization: 18.1% throughput improvement, 10.7% TTFT improvement on DeepSeek-V3.1 (#29345)
  • Shared Experts Overlap with FlashInfer DeepGEMM: 2.2% throughput improvement, 3.6% TTFT improvement at batch size 32 (#28879)
  • DeepGEMM N dim restriction reduced from 128 to 64 multiplier (#28687)
  • DeepEP low-latency with round-robin expert placement (#28449)
  • NVFP4 MoE CUTLASS support for SM120 (#29242)
  • H200 Fused MoE Config improvements (#28992)

• AMD ROCm:

  • DeepSeek v3.2 and SparseMLA support (#26670)
  • FP8 MLA decode support (#28032)
  • AITER sampling ops integration (#26084)
  • AITER triton attention backend (#28701)
  • Bitsandbytes quantization on AMD GPUs with warp size 32 (#27307)
  • Fastsafetensors support (#28225)
  • Sliding window support for AiterFlashAttentionBackend (#29234)
  • Whisper v1 with Aiter Unified/Flash Attention (#28376)

• CPU:

  • Paged attention GEMM acceleration on ARM CPUs with NEON (#29193)
  • Parallelize over tokens in int4 MoE (#29600)
  • CPU all reduce optimization for async_scheduling + DP>1 (#29311)

• Attention: FlashAttention ViT support, now default backend (#28763).

• Long Context: Optimized gather_and_maybe_dequant_cache kernel for extremely long sequences (#28029).

• Multi-NUMA: Enhanced NUMA functionality for systems with multiple NUMA nodes per socket (#25559).

• Docker: Image size reduced by ~200MB (#29060).

Quantization

• W4A8: Marlin kernel support (#24722).
• NVFP4:
  • MoE CUTLASS support for SM120 (#29242)
  • TRTLLM MoE NVFP4 kernel (#28892)
  • CuteDSL MoE with NVFP4 DeepEP dispatch (#27141)
  • Non-gated activations support in modelopt path (#29004)
• AWQ: Compressed-tensors AWQ support for Turing GPUs (#29732).
• LoRA: FusedMoE LoRA Triton kernel for MXFP4 (#29708).
• Online quantization: Moved to model.load_weights (#26327).

API & Frontend

• Responses API:
  • Multi-turn support for non-harmony requests (#29175)
  • Reasoning item input parsing (#28248)
• Tool Calling:
  • Parsed tool arguments support (#28820)
  • parallel_tool_calls param compliance (#26233)
  • Tool filtering support in ToolServer (#29224)
• Whisper: verbose_json and timestamp features for transcription/translation (#24209).
• Sampling: Flat logprob control moved from env var to SamplingParams (#28914).
• GGUF: Improved HuggingFace loading UX with repo_id:quant_type syntax (#29137).
• Profiling: Iteration-level profiling for Torch and CUDA profiler (#28987).
• Logs: Colorized log output (#29017).
• Optimization Levels: -O0, -O1, -O2, -O3 allow trading startup time for performance, more compilation flags will be added in future releases (#26847)

Dependencies

• PyTorch 2.9.0 with CUDA 12.9 (#24994) - Breaking change requiring environment updates.
• xgrammar: Updated to 0.1.27 (#28221).
• Transformers: Updated to 4.57.3 (#29418), preparation for v5 with rope_parameters (#28542).
• XPU: torch & IPEX 2.9 upgrade (#29307).

V0 Deprecation & Breaking Changes

Removed Parameters:

• num_lookahead_slots (#29000)
• best_of (#29090)
• LoRA extra vocab (#28545)

Deprecated:

• xformers backend (#29262)
• seed=None (#29185)

Scheduled Removals (will be removed in future release):

• ParallelConfig's direct child EPLB fields (#29324)
• guided_* config fields (#29326)
• override_pooler_config and disable_log_requests (#29402)
• CompilationConfig.use_inductor (#29323)
• Deprecated metrics (#29330)

Other Breaking Changes:

• PyTorch 2.9.0 upgrade requires CUDA 12.9 environment
• Mistral format auto-detection for model loading (#28659)

New Contributors

• @jesse996 made their first contribution in https://github.com/vllm-project/vllm/pull/28846
• @Nepherpitou made their first contribution in https://github.com/vllm-project/vllm/pull/28960
• @Samoed made their first contribution in https://github.com/vllm-project/vllm/pull/27329
• @j20120307 made their first contribution in https://github.com/vllm-project/vllm/pull/28999
• @vnadathur made their first contribution in https://github.com/vllm-project/vllm/pull/26468
• @zhyajie made their first contribution in https://github.com/vllm-project/vllm/pull/28942
• @IzzyPutterman made their first contribution in https://github.com/vllm-project/vllm/pull/28896
• @rjrock-amd made their first contribution in https://github.com/vllm-project/vllm/pull/28905
• @zq1997 made their first contribution in https://github.com/vllm-project/vllm/pull/27715
• @shengliangxu made their first contribution in https://github.com/vllm-project/vllm/pull/28076
• @prashanth058 made their first contribution in https://github.com/vllm-project/vllm/pull/28972
• @qgallouedec made their first contribution in https://github.com/vllm-project/vllm/pull/28820
• @zhanggzh made their first contribution in https://github.com/vllm-project/vllm/pull/19347
• @pandalee99 made their first contribution in https://github.com/vllm-project/vllm/pull/26628
• @dsuhinin made their first contribution in https://github.com/vllm-project/vllm/pull/29100
• @xli made their first contribution in https://github.com/vllm-project/vllm/pull/29124
• @jeremyteboul made their first contribution in https://github.com/vllm-project/vllm/pull/29059
• @soodoshll made their first contribution in https://github.com/vllm-project/vllm/pull/28875
• @bhagyashrigai made their first contribution in https://github.com/vllm-project/vllm/pull/28957
• @skaraban3807 made their first contribution in https://github.com/vllm-project/vllm/pull/25559
• @Victor49152 made their first contribution in https://github.com/vllm-project/vllm/pull/28892
• @rjrock made their first contribution in https://github.com/vllm-project/vllm/pull/29205
• @FlintyLemming made their first contribution in https://github.com/vllm-project/vllm/pull/29182
• @madskildegaard made their first contribution in https://github.com/vllm-project/vllm/pull/29175
• @nandan2003 made their first contribution in https://github.com/vllm-project/vllm/pull/29189
• @michaelact made their first contribution in https://github.com/vllm-project/vllm/pull/29173
• @yongming-qin made their first contribution in https://github.com/vllm-project/vllm/pull/28958
• @joshiemoore made their first contribution in https://github.com/vllm-project/vllm/pull/29249

Full Changelog: https://github.com/vllm-project/vllm/compare/v0.11.1...v0.12.0