Qwen3.5-35B-A3B
Qwen · Open weights · Released Feb 25, 2026 · qwen/qwen3.5-35b-a3b
Best $0.08 in / $0.75 out via Darkbloom · fp4
262K contextVisionToolsReasoningStructured outputPrompt cachingOpen weights
Available from 7 providersUSD / MTOK · PRICES 23 MINUTES AGO · SPEED 10 MINUTES AGO
EST. MONTHLY COSTclick a row for detail · click headers to sort
| PROVIDER | INPUT | OUTPUT | CACHE | CONTEXT | QUANT | TTFT | TPS | SMOKE | UPTIME |
|---|---|---|---|---|---|---|---|---|---|
 Darkbloom | $0.08 | $0.75 | $0.04 | 262K | fp4 | 3.5s° | 108° tok/s | — | 100.0% |
 DeepInfra | $0.14 | $1.00 | $0.05 | 262K | fp8 | 1.8s° | 104° tok/s | 3/5 | 99.8% |
 Parasail | $0.15 | $1.00 | $0.05 | 262K | fp8 | ✕ | ✕ | 3/5 | 100.0% |
 Alibaba | $0.163 | $1.30 | — | 262K | — | 1.5s° | 202° tok/s | 5/5 | 95.5% |
 AtlasCloud | $0.225 | $1.80 | $0.225 | 262K | fp8 | ✕ | ✕ | 5/5 | 100.0% |
 SiliconFlow | $0.24 | $1.80 | $0.15 | 262K | fp8 | ✕ | ✕ | 3/5 | 86.2% |
 Venice | $0.313 | $1.25 | $0.156 | 256K | — | 1.3s° | 221° tok/s | — | 99.8% |
TTFT/TPS: median (P50) over trailing 504h · chat10k profile (~10K in / 200 out) · measured against the provider API directly; ° = routed via OpenRouter (adds a hop) · ✕ = probed, endpoint returned no measurable stream · SMOKE: of 5 programmatic capability probes passed (JSON schema, instruction following, tool call, long-context, benign compliance; retry-once) · screened via OpenRouter routing · v0.2 · methodology v1.1 · measured 10 minutes ago
SERVE THIS MODEL AND NOT LISTED? LIST YOUR INFERENCE →
Pricing variants · USD / MTok in / out
STANDARD
$0.08 / $0.75
per MTok in / out
CACHE READ $0.04
Recent changesRSS ↗
Oct 6, 2026endpoint removed · cheaperinference8 → 7
Sep 23, 2026endpoint added · cheaperinference7 → 8
Sep 5, 2026endpoint removed · coreweave8 → 7
Aug 27, 2026endpoint added · darkbloom7 → 8
About
The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall...
Specifications
Context window262,144
Max output235,929
Modalities intext, image, video
Modalities outtext
LicenseOpen weights
ReleasedFeb 25, 2026
TokenizerQwen3
WeightsHugging Face ↗
Related
Embed badge
[](https://modelindex.ai/models/qwen/qwen3.5-35b-a3b)Sources
OpenRouter API23 minutes ago
AGGREGATORHugging Face23 minutes ago
COMMUNITYModelIndex benchmarks18 days ago
MEASUREDVerified 23 minutes ago · list prices, not negotiated rates.