LLM Reference

Qwen3.5-35B-A3B vs Qwen3.5-397B-A17B

Qwen3.5-35B-A3B (2026) and Qwen3.5-397B-A17B (2026) are frontier-tier reasoning models from Alibaba. Qwen3.5-35B-A3B ships a 262k-token context window, while Qwen3.5-397B-A17B ships a 262k-token context window. On MMLU PRO, Qwen3.5-397B-A17B leads by 2.5 pts. On pricing, Qwen3.5-35B-A3B costs $0.14/1M input tokens versus $0.39/1M for the alternative. This comparison covers specs, pricing, API access, capabilities, benchmarks, input and output token costs, and production fit for coding and agent workloads.

Qwen3.5-35B-A3B is ~181% cheaper at $0.14/1M; pay for Qwen3.5-397B-A17B only for vision-heavy evaluation.

Decision scorecard

Local evidence first
SignalQwen3.5-35B-A3BQwen3.5-397B-A17B
Best forreasoning-heavy apps, tool-calling agents, and provider-routed productionreasoning-heavy apps, multimodal apps, and tool-calling agents
Decision fitCoding, RAG, and AgentsCoding, RAG, and Agents
Context window262k262k
Cheapest output$1/1M tokens$2.34/1M tokens
Provider routes2 tracked4 tracked
Shared benchmarks7 sharedMMLU PRO leader

Decision tradeoffs

Choose Qwen3.5-35B-A3B when...
  • Qwen3.5-35B-A3B has the lower cheapest tracked output price at $1/1M tokens.
  • Local decision data tags Qwen3.5-35B-A3B for Coding, RAG, and Agents.
Choose Qwen3.5-397B-A17B when...
  • Qwen3.5-397B-A17B holds a shared-benchmark lead on MMLU PRO, ahead by 2.5 points.
  • Qwen3.5-397B-A17B has broader tracked provider coverage for fallback and route flexibility.
  • Qwen3.5-397B-A17B uniquely exposes Vision and Multimodal in local model data.
  • Local decision data tags Qwen3.5-397B-A17B for Coding, RAG, and Agents.

Monthly cost at traffic

Estimate token spend from the cheapest tracked input and output route or tier on this page.

Lower estimate Qwen3.5-35B-A3B

Qwen3.5-35B-A3B

$361

Cheapest tracked route/tier: OpenRouter

Qwen3.5-397B-A17B

$897

Cheapest tracked route/tier: OpenRouter

Estimated monthly gap: $536. Batch, cache, alternate speed tiers, and negotiated pricing are excluded from this local estimate.

Switch friction

Qwen3.5-35B-A3B -> Qwen3.5-397B-A17B
  • Provider overlap exists on OpenRouter and Novita AI; start route-level A/B tests there.
  • Qwen3.5-397B-A17B is $1.34/1M tokens higher on cheapest tracked output pricing, so quality gains need to justify the spend.
  • Qwen3.5-397B-A17B adds Vision and Multimodal in local capability data.
Qwen3.5-397B-A17B -> Qwen3.5-35B-A3B
  • Provider overlap exists on OpenRouter and Novita AI; start route-level A/B tests there.
  • Qwen3.5-35B-A3B is $1.34/1M tokens lower on cheapest tracked output pricing before cache, batch, or negotiated discounts.
  • Check replacement coverage for Vision and Multimodal before moving production traffic.

Specs

Specification
Released2026-02-242026-02-16
Context window262k262k
Parameters35B397B
ArchitectureMixture of ExpertsMixture of Experts
LicenseApache 2.0OSI-approvedApache 2.0OSI-approved
OpennessOpen sourceOpen source
WeightsAvailableAvailable
CodeUnknownUnknown
Commercial useCommercial use: permittedCommercial use: permitted
Knowledge cutoff--

Pricing and availability

Pricing attributeQwen3.5-35B-A3BQwen3.5-397B-A17B
Input price$0.14/1M tokens$0.39/1M tokens
Output price$1/1M tokens$2.34/1M tokens
Providers

Capabilities

CapabilityQwen3.5-35B-A3BQwen3.5-397B-A17B
VisionNoYes
MultimodalNoYes
ReasoningYesYes
JSON / Tool useYesYes
Structured outputsYesYes
Code executionNoNo
IDE integrationNoNo
Computer useNoNo
Parallel agentsNoNo

Benchmarks

BenchmarkQwen3.5-35B-A3BQwen3.5-397B-A17B
MMLU PRO85.387.8
SWE-bench Verified69.276.2
Google-Proof Q&A84.589.3
LiveCodeBench74.683.6
Humanity's Last Exam22.428.7
Berkeley Function Calling Leaderboard v367.372.9
τ-bench81.286.7

Continue comparing

Last reviewed: 2026-06-29. Data sourced from public model cards and provider documentation.