chore(deps): bump flashinfer-python from 0.6.12 to 0.6.14

[dependabot]

Bumps flashinfer-python from 0.6.12 to 0.6.14.

Release notes

Sourced from flashinfer-python's releases.

Release v0.6.14

Manual Intervention Needed ⚠️

Please modify your install instructions as you bump the version to 0.6.14.

pip install flashinfer-python
pip install flashinfer-cubin --index-url https://flashinfer.ai/whl     # this is the difference
pip install flashinfer-jit-cache --index-url https://flashinfer.ai/whl/cu129
OR
pip install flashinfer-jit-cache --index-url https://flashinfer.ai/whl/cu130

See also flashinfer-ai/flashinfer#3808

Highlights

This release pushes FlashInfer's coverage onto Blackwell RTX PRO and DGX Spark silicon, lands the CuTe-DSL rewrite of the gated-delta-rule (GDN) kernels behind Qwen3.5/3.6, and makes those kernels production-ready inside vLLM.

Gemma 4 and W4A16 on Blackwell RTX Pro and DGX Spark

Blackwell SM12x parts (RTX PRO 6000, DGX Spark / GB10) previously trailed Blackwell GB200 NVL72 (and B300 SM103) on attention shape and quantization coverage. This release closes the gap for Gemma 4 and weight-only-quantized inference: head_dim=512 attention for Gemma 4's global layers now runs on SM120/121, the FMHAv2 prefill path gains head_dim=256/512 and sliding-window masking on SM120, the new mm_bf16_fp4 W4A16 GEMM is tuned for DGX Spark (completing W4A16 across both dense GEMM and MoE), and gated tanh-GELU brings Gemma 4 MoE onto the CUTLASS backend.

• flashinfer-ai/flashinfer#3576
• flashinfer-ai/flashinfer#3518
• flashinfer-ai/flashinfer#3597
• flashinfer-ai/flashinfer#3501

GDN / gated delta rule: CuTe-DSL overhaul (Qwen3.5/3.6)

The gated-delta-rule kernels behind the Qwen3.5/3.6 family were rewritten from CUTLASS C++ to CuTe-DSL (flashinfer-ai/flashinfer#3491). The rewrite eliminates the C++ JIT compilation pain reported by customers and establishes the base for context-parallel delta-rule kernels — covering SM90 prefill and its context-parallel variant, SM120 prefill, and a ~20–25% GDN prefill speedup from mainloop efficiency work.

• flashinfer-ai/flashinfer#3477
• flashinfer-ai/flashinfer#3481
• flashinfer-ai/flashinfer#3479
• flashinfer-ai/flashinfer#3742

GDN production-ready in vLLM

Two gaps blocked GDN serving in vLLM (flashinfer-ai/flashinfer#3602); both are now resolved. GDN kernels are compilation batch-size agnostic, so a single compiled cubin is reused across batch shapes instead of recompiling on every new batch size at inference time, and a new BF16 state recovery/decode kernel writes SSM state into preallocated space to supply the MTP-compatible spec-decode path vLLM needs.

• flashinfer-ai/flashinfer#3649
• flashinfer-ai/flashinfer#3502

DeepSeek-class sparse MLA on Blackwell, FP8 KV on Hopper

New sparse-MLA paged-attention kernels extend the DeepSeek-V4 (d_qk=512) and DeepSeek-V3.2 / GLM-5.1 (d_qk=576) families onto SM120/121 through the existing flashinfer.mla APIs, with DSv4 coverage broadened to 8/16/32 head counts. On Hopper SM90, native FP8 KV cache support eliminates SGLang's per-layer cast workaround, saving an HBM round-trip per layer on DeepSeek-V3/V4 while staying bit-identical to the BF16 path.

... (truncated)

Commits

• 19f1a41 update version to 0.6.14
• f959730 fix: remove @​flashinfer_api from moe_gemm_mxfp8_nt_groupwise to defer API com...
• be65818 fix(aot): exclude head_dim=512 FA2 modules from AOT jit-cache wheel (#3769)
• 1657d13 docs: add missing arg docs for output_scalar_scale and activation (#3749)
• 4fb217e create rc1
• 15a2459 Remove pagesize 16/32 assertion from xqa nvfp4 sm120 (#3724)
• 5f2bdc4 test: fix CUDA OOM in batch-prefill custom-mask test on 24GB CI GPUs (#3609)
• 2d4db82 Improve GDN prefill perf by ~20-25% (#3742)
• c35fb99 feat(moe): enable DSFp8 + LoRA delta path (#3708)
• 2099bac docs(gdn): document missing ssm_state_indices param in gated_delta_rule_mtp (...
• Additional commits viewable in compare view

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[dependabot]

OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting @dependabot ignore this major version or @dependabot ignore this minor version. You can also ignore all major, minor, or patch releases for a dependency by adding an ignore condition with the desired update_types to your config file.

If you change your mind, just re-open this PR and I'll resolve any conflicts on it.