Qwen3.8-Flash-Next vs Step 5 Preview
Alibaba · China | StepFun · China · Updated June 2026
Quick verdict
Pick Qwen3.8-Flash-Next for swe-bench pro (62.5, ahead of claude opus 4.6 max's 53.4) or cost efficiency: ~1/9th the training cost of qwen3.7-plus, ~12x cheaper api than flagship qwen3.8-max. Pick Step 5 Preview for sparse moe architecture — 600b total parameters, only 27b active per token — built for long-horizon agentic work across large codebases and documents or 1m-token context window (64k max output) at $1/$2.70 per million tokens, among the cheapest frontier-scale context windows available. Choose Qwen3.8-Flash-Next if you need self-hosting or data privacy; Step 5 Preview if you want a managed API.
Qwen3.8-Flash-Next (Alibaba) and Step 5 Preview (StepFun) are two of the models people most often weigh against each other in 2026. Qwen3.8-Flash-Next is alibaba's open-weight preview of its next-generation Qwen4 hybrid-attention architecture, released August 26, 2026 as a 125B-parameter (6B active) MoE model priced roughly 12x below its own flagship Qwen3.8-Max. Step 5 Preview is stepFun's September 20, 2026 flagship preview — a 600B-parameter sparse MoE model (27B active) with a 1M-token context window at $1/$2.70 per million tokens, open weights promised for mid-October. They diverge most on price, context window and open vs. closed weights — each quantified below from the models' real specs.
Key differences at a glance
- ▸Price: Qwen3.8-Flash-Next is about 6.3× cheaper on input ($0.16/$0.47 per 1M tokens vs $1/$2.7 per 1M tokens) — a large enough gap that at scale it can be the single biggest line item in the decision.
- ▸Context window: Step 5 Preview holds 3.8× more — 1M tokens (~1,500 pages) vs 262K tokens natively (extensible to 1M with YaRN) (~393 pages). But effective recall usually fades long before the advertised ceiling, so the bigger number only helps if the model reasons over it.
- ▸Recency: Step 5 Preview is the newer model by about 25 days (released September 20, 2026), usually meaning fresher training data and capabilities.
Side-by-side specs
| Spec | Qwen3.8-Flash-Next | Step 5 Preview |
|---|---|---|
| Provider | Alibaba (China) | StepFun (China) |
| Released | August 26, 2026 | September 20, 2026 |
| Context window | 262K tokens natively (extensible to 1M with YaRN) (~393 pages) | 1M tokens (~1,500 pages) |
| Price (in/out) | $0.16/$0.47 per 1M tokens | $1/$2.7 per 1M tokens |
| Open weight? | Yes — self-hostable | No — API only |
| Modalities | text, image, video | text, image, code |
| SWE-Bench Verified | Not published | Not published |
| MRCR v2 @ 1M | Not published | Not published |
Who wins what
SWE-bench Pro (62.5, ahead of Claude Opus 4.6 Max's 53.4)
Qwen3.8-Flash-Next
Step 5 Preview is comparatively weak here — trails closed frontier models on coding — 67.7% on DeepSWE v1.1 vs GPT-6 Astra's 74.1% and Claude Opus 5's 74.0%
Cost efficiency: ~1/9th the training cost of Qwen3.7-Plus, ~12x cheaper API than flagship Qwen3.8-Max
Qwen3.8-Flash-Next
At $0.16/$0.47 per 1M tokens it undercuts Step 5 Preview ($1/$2.7 per 1M tokens), and that gap compounds at volume.
Vision-based agentic tasks (AndroidWorld: 84.5)
Qwen3.8-Flash-Next
Alibaba's open-weight preview of its next-generation Qwen4 hybrid-attention architecture, released August 26, 2026 as a 125B-parameter (6B active) MoE model priced roughly 12x below its own flagship Qwen3.8-Max — and it runs cheaper at $0.16/$0.47 per 1M tokens.
Sparse MoE architecture — 600B total parameters, only 27B active per token — built for long-horizon agentic work across large codebases and documents
Step 5 Preview
Its 1M tokens window holds about 3.8× more than Qwen3.8-Flash-Next's 262K tokens natively (extensible to 1M with YaRN) in a single prompt.
1M-token context window (64K max output) at $1/$2.70 per million tokens, among the cheapest frontier-scale context windows available
Step 5 Preview
StepFun's September 20, 2026 flagship preview — a 600B-parameter sparse MoE model (27B active) with a 1M-token context window at $1/$2.70 per million tokens, open weights promised for mid-October — and it carries the larger 1M tokens context.
67.7% on DeepSWE v1.1, ahead of open-weight rivals Kimi K3 (67.5%) and GLM-5.3 (66.9%)
Step 5 Preview
Qwen3.8-Flash-Next is comparatively weak here — an open-weight architecture preview rather than Alibaba's polished flagship product
Lowest cost at scale
Qwen3.8-Flash-Next
At $0.16/$0.47 per 1M tokens, it is the cheaper of the two — the gap dominates the bill on high-volume workloads.
Largest single-prompt input
Step 5 Preview
Its 1M tokens window is about 3.8× larger than Qwen3.8-Flash-Next's 262K tokens natively (extensible to 1M with YaRN), fitting roughly 1,500 pages in one prompt.
Which should you pick?
A cost-sensitive startup shipping high volume
→ Qwen3.8-Flash-Next
At $0.16/$0.47 per 1M tokens it undercuts Step 5 Preview, and on millions of tokens that margin decides the monthly bill.
Someone analysing very long documents or codebases
→ Step 5 Preview
Larger 1M tokens window fits more in one prompt.
A team with data-privacy or self-hosting needs
→ Qwen3.8-Flash-Next
Open weights let you run it on your own hardware; Step 5 Preview is API-only.
Anyone whose priority is swe-bench pro (62.5, ahead of claude opus 4.6 max's 53.4)
→ Qwen3.8-Flash-Next
It is specifically built for that.
Anyone whose priority is sparse moe architecture — 600b total parameters, only 27b active per token — built for long-horizon agentic work across large codebases and documents
→ Step 5 Preview
That is its strongest area.
Qwen3.8-Flash-Next: where it fits
Alibaba's open-weight preview of its next-generation Qwen4 hybrid-attention architecture, released August 26, 2026 as a 125B-parameter (6B active) MoE model priced roughly 12x below its own flagship Qwen3.8-Max. Released August 26, 2026 by Alibaba, it is built for sWE-bench Pro (62.5, ahead of Claude Opus 4.6 Max's 53.4), cost efficiency: ~1/9th the training cost of Qwen3.7-Plus, ~12x cheaper API than flagship Qwen3.8-Max, vision-based agentic tasks (AndroidWorld: 84.5), and previews Qwen4's hybrid gated-DeltaNet plus sparse-attention architecture.
Its trade-offs are real: trails Claude Opus 4.6 Max on Humanity's Last Exam (35.9 vs 40.0), lower OSWorld 2.0 binary success rate (19.4%), and an open-weight architecture preview rather than Alibaba's polished flagship product. At $0.16 in / $0.47 out per million tokens, it sits in the budget price band.
Step 5 Preview: where it fits
StepFun's September 20, 2026 flagship preview — a 600B-parameter sparse MoE model (27B active) with a 1M-token context window at $1/$2.70 per million tokens, open weights promised for mid-October. Released September 20, 2026 by StepFun, it is built for sparse MoE architecture — 600B total parameters, only 27B active per token — built for long-horizon agentic work across large codebases and documents, 1M-token context window (64K max output) at $1/$2.70 per million tokens, among the cheapest frontier-scale context windows available, 67.7% on DeepSWE v1.1, ahead of open-weight rivals Kimi K3 (67.5%) and GLM-5.3 (66.9%), and scores 44 on the Artificial Analysis Intelligence Index, ranking 27th of 653 models tracked.
Its trade-offs: trails closed frontier models on coding — 67.7% on DeepSWE v1.1 vs GPT-6 Astra's 74.1% and Claude Opus 5's 74.0%, open weights were promised for October 15, 2026 but hadn't shipped as of this writing — API-only for now despite the open-weight roadmap, a 'Preview' release, not StepFun's finished flagship — positioning and pricing may change once weights ship, and stepCodeBench and DeepSWE figures are StepFun's own reported numbers, not yet independently reproduced at scale. At $1 in / $2.7 out per million tokens, it sits in the budget price band.
The bottom line for this matchup
The defining split here is open vs. closed. Qwen3.8-Flash-Next gives you weights you control — self-host it, fine-tune it, keep data in-house, pay only for hardware. Step 5 Preview gives you a managed, always-updated API with no infrastructure to run. Teams with GPUs, privacy requirements, or huge volume often favour the open model; teams that want zero ops and the latest capabilities favour the closed one. Capability is close enough that this operational question, not the benchmark, usually decides it.
Want both Qwen3.8-Flash-Next and Step 5 Preview without two subscriptions? LumiChats gives you these plus 40+ models under one ₹69/day pass (about $1/day) — draft with one, cross-check with the other.
See pricingFrequently asked questions
Is Qwen3.8-Flash-Next or Step 5 Preview better for coding?
Public SWE-Bench figures are not available for either model, so the honest test is your own repository — run an identical real bug through both. By design, Qwen3.8-Flash-Next leans toward swe-bench pro (62.5, ahead of claude opus 4.6 max's 53.4) while Step 5 Preview leans toward sparse moe architecture — 600b total parameters, only 27b active per token — built for long-horizon agentic work across large codebases and documents, and that positioning usually predicts which feels better on your codebase.
Which is cheaper, Qwen3.8-Flash-Next or Step 5 Preview?
Qwen3.8-Flash-Next is open-weight, so self-hosting means no per-token fee (you pay for hardware instead), while Step 5 Preview is API-metered at $1/$2.7 per 1M tokens. For most teams without GPUs, the API model is cheaper to start; at very high volume, self-hosting can win.
Which has the bigger context window?
Step 5 Preview — 1M tokens vs 262K tokens natively (extensible to 1M with YaRN), about 3.8× larger. Useful only if the model actually reasons over the full window, which not all do.
Can I use both Qwen3.8-Flash-Next and Step 5 Preview together?
Yes — a multi-model platform like LumiChats gives you Qwen3.8-Flash-Next, Step 5 Preview and 40+ others under one ₹69/day pass (about $1/day), so you can draft with one and cross-check with the other instead of buying two subscriptions.
Which is newer, Qwen3.8-Flash-Next or Step 5 Preview?
Step 5 Preview — released September 20, 2026, about 25 days after Qwen3.8-Flash-Next.
Related comparisons
Specifications and benchmarks reflect publicly reported figures as of June 2026 and may change as providers release updates. Always verify on your own workload.