MODEL PROFILE

Qwen3.5-122B-A10B-int4-fp8-hybrid

bleysg · Qwen3.5-122B-A10B-int4-fp8-hybrid

A hybrid-quantized checkpoint of **Qwen3.5-122B-A10B** for bandwidth-optimal decode on the **NVIDIA DGX Spark (GB10 / SM121, 128 GB unified)** under vLLM. Routed experts and attention stay INT4; the BF16 **shared experts** (dense — read on every token) are replaced with **calibrated FP8** from the official FP8 release. - **Routed experts + attention:** INT4 (GPTQ / AutoRound) — from `Intel/Qwen3.5-122B-A10B-int4-AutoRound`. - **Shared experts:** FP8 E4M3 block-128 — from `Qwen/Qwen3.5-122B-A10B-FP8`. - Embeddings / norms / head: unchanged from the INT4 base. 144 shared-expert layers convert BF16→FP8 — an always-on bandwidth lever worth **+28 % at base / no-spec decode** on GB10 (28.2 → 36.0 tok/s), diminishing under high-acceptance speculative decode (see the project's amortization law).

Text generationlicense:other
Context windowNot disclosed
Input → OutputText → Text
Model weightsNot verified
Profile updated2026/10/29 19:12:24
OVERVIEW

Model overview

A hybrid-quantized checkpoint of **Qwen3.5-122B-A10B** for bandwidth-optimal decode on the **NVIDIA DGX Spark (GB10 / SM121, 128 GB unified)** under vLLM. Routed experts and attention stay INT4; the BF16 **shared experts** (dense — read on every token) are replaced with **calibrated FP8** from the official FP8 release. - **Routed experts + attention:** INT4 (GPTQ / AutoRound) — from `Intel/Qwen3.5-122B-A10B-int4-AutoRound`. - **Shared experts:** FP8 E4M3 block-128 — from `Qwen/Qwen3.5-122B-A10B-FP8`. - Embeddings / norms / head: unchanged from the INT4 base. 144 shared-expert layers convert BF16→FP8 — an always-on bandwidth lever worth **+28 % at base / no-spec decode** on GB10 (28.2 → 36.0 tok/s), diminishing under high-acceptance speculative decode (see the project's amortization law).

Key features

特点摘要待依据来源补齐,不凭模型名称推断能力。

Limitations & considerations

Sources are insufficient to establish specific limitations. Verify task performance, latency, data policies, licensing and version lifecycle.

PRICING & COST

Source prices & cost analysis

Prices vary by channel, region and configuration. Channel quotes are listed independently and are not treated as a single official provider price.

Source / channelInputOutputBilling unitCollected at
No verified prices yet. Check the source channel.

Estimate monthly base token costs

—

Estimates exclude cache discounts, tiered pricing, multimodal surcharges, web search, taxes and additional reasoning output.

INDEPENDENT BENCHMARKS

Public benchmarks & run configuration

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

View full indexed rankings →

Frequently asked questions

Do open weights permit free commercial use?

Not necessarily. Check the version license, commercial conditions and usage policies. Self-hosting incurs compute and maintenance costs.

Are these real-time billing prices?

These are timestamped source snapshots, not live billing prices. Check the channel for current versions, regions, discounts and billing methods.

Sources & traceability

许可证:other。模型介绍与第三方评测均以原来源的具体版本和方法为准。