ModelsQwen3-VL-8B Thinking: Paying for Reasoning on Images
The Thinking edition of Qwen3-VL-8B reasons before it answers. Which scores that lifts, which it leaves flat, and why the cost sits almost entirely in output tokens.
ReadNotes on models, agents, and what inference costs
Practical writing for developers building with large language models — how they work, how to pick one, and how to keep the bill predictable.
Picking between models — and testing them on your own work.
Models
ModelsQwen3.5-122B-A10B: The Agentic Pick of the Medium Qwen3.5s
Qwen3.5-122B-A10B pairs hybrid linear attention with 256 experts and 10B active. Where its size shows in the benchmarks, where it does not, and what serving it takes.
Read
ModelsQwen3.5-27B: The Dense Model That Keeps Up With Bigger Siblings
Qwen3.5-27B is the only dense model in the medium Qwen3.5 release and matches the 122B on several coding tests. Why dense matters, what it costs, and when to move to 3.6.
Read
ModelsQwen3.5-35B-A3B: 3B Active Parameters, Medium-Model Scores
Qwen3.5-35B-A3B runs only 3B parameters per token yet tracks its larger siblings on many benchmarks. Where it is strong, where it is thin, and how to use it as a workhorse.
ReadModelsQwen3.5-397B-A17B: The Open Flagship Behind Qwen3.5-Plus
A 397B MoE with 17B active, hybrid linear attention, vision input and Apache 2.0 weights. What the architecture buys you and when to self-host instead of calling Plus.
ReadModelsQwen3.5-9B Guide: A Small Multimodal Model Worth Measuring
Qwen3.5-9B brings image and video input, a 262K window and Apache 2.0 weights to a size you can run locally. Where it holds up against its big siblings and where it does not.
Read
ModelsQwen3.5-Flash: Hosted 35B-A3B With a Flat-Priced 1M Window
Qwen3.5-Flash is the hosted Qwen3.5-35B-A3B with a 1M default window and one price across all of it. What you are calling, what it costs and when to pick it over 3.6 Flash.
ReadModelsQwen3.5-Plus: Two Snapshots, a 1M Window and a 256K Price Step
Qwen3.5-Plus is the hosted Qwen3.5-397B-A17B with a 1M window and built-in tools. How its two snapshots differ in practice, where the price steps up, and when to choose it.
Read
ModelsQwen3.6 Plus Guide: Is the April Release Still Worth Calling?
Qwen3.6 Plus brought a 1M window and video input to Alibaba's hosted mid tier in April 2026. What it is, what it costs past 256K, and when to move to 3.7 Plus.
ReadModelsQwen3.6-35B-A3B: What Changed in Qwen's 3B-Active Model
Qwen3.6-35B-A3B keeps the size of its predecessor and moves its gains into agentic coding. What improved, what slipped, and how it compares with the dense Qwen3.6 27B.
ReadModelsQwen3.6-Flash: Pricing, Limits and the Open Model Behind It
Qwen3.6-Flash is the hosted Qwen3.6-35B-A3B, with a 1M window and tiered pricing well above 3.5 Flash. What is documented, what it costs and how to decide.
Read
ModelsQwen3.7 Flash: Read the Tiered Price Before You Commit
Qwen3.7 Flash is the cheapest vision model in the Qwen3.7 line, but its price steps up twice as prompts grow. Where the tiers fall, what it is for, and when to use Plus.
ReadShowing 277–288 of 446 articles