Qwen3.8-Flash-Next vs Step 5 Preview
Full profiles: Qwen3.8-Flash-Next (Alibaba) and Step 5 Preview (StepFun): every benchmark, price and rank.
Updated · Data from 7 sources
As of 6 Nov 2026, Qwen3.8-Flash-Next and Step 5 Preview are neck and neck on quality (54.6 vs 54.9 on the BenchLeader Index, inside the ±10.4 margin), and Qwen3.8-Flash-Next is 6.2× cheaper at $0.230 vs $1.43 per 1M tokens blended. Choose Qwen3.8-Flash-Next for coding (66 vs 50), agents & tools (64 vs 51), and multimodal (62 vs 59); choose Step 5 Preview for fast first answers (25.9 s vs 125.1 s) and knowledge (55 vs 38), reasoning (53 vs 45), and long context (69 vs 65), where it also streams 3×+ faster.
Key facts (as of 6 Nov 2026)
- Qwen3.8-Flash-Next scores 54.6 on the BenchLeader Index and Step 5 Preview scores 54.9 (BenchLeader, 6 Nov 2026).
- Qwen3.8-Flash-Next costs $0.230 per 1M tokens blended (list price, 6 Nov 2026).
- Step 5 Preview costs $1.43 per 1M tokens blended (list price, 6 Nov 2026).
- Step 5 Preview streams 3×+ faster than Qwen3.8-Flash-Next (Artificial Analysis, 7-day median, 6 Nov 2026).
- Of the 18 benchmarks both models have, Step 5 Preview leads 10 (BenchLeader, 6 Nov 2026).
Qwen3.8-Flash-Next benchmarks 
Step 5 Preview benchmarks Which to pick for…
The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.
| Use case | Pick | Why | Confidence |
|---|---|---|---|
| Coding | Qwen3.8-Flash-Next | 66 vs 50 category score; SciCode (AA) 50.6% vs 58.9% | High |
| Agents & tool use | Qwen3.8-Flash-Next | 64 vs 51 category score; Arena Agent -4.4 vs -2.9 | High |
| Reasoning | Step 5 Preview | 53 vs 45 category score; HLE-Diamond (AA) 24.3% vs 16.3% | High |
| Knowledge | Step 5 Preview | 55 vs 38 category score; AA-Omniscience v1.1 16.4 vs -12.2 | High |
| Long documents | Step 5 Preview | 69 vs 65 category score; AA-LCR v1.1 88.3% vs 79.7% | Medium |
| Images and other media | Qwen3.8-Flash-Next | 62 vs 59 category score; MMMU-Pro 79.8% vs 76.4% | Medium |
| Chat quality (human preference) | – | Only Step 5 Preview has an Arena rating (63 category score) | Not comparable |
| Lowest cost | Qwen3.8-Flash-Next | $0.230 vs $1.43 per 1M blended; agentic coding session $0.011 vs $0.071 | – |
| Fastest first answer | Step 5 Preview | 25.9 s (Step 5 Preview) vs 125.1 s (Qwen3.8-Flash-Next) to the first answer token | – |
| High volume on a budget | Neither flagship | A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs | – |
Key differences
- Qwen3.8-Flash-Next and Step 5 Preview are neck and neck on quality (54.6 vs 54.9, within the ±10.4 margin).
- Qwen3.8-Flash-Next is stronger in agents & tools, coding, multimodal.
- Step 5 Preview is stronger in knowledge, long context, reasoning, composite, human preference, maths.
- Qwen3.8-Flash-Next is 6.2× cheaper ($0.230 vs $1.43 per 1M blended).
- Step 5 Preview streams 2.1× faster (132 vs 64 tokens per second).
At a glance
| Spec | Qwen3.8-Flash-Next | Step 5 Preview |
|---|---|---|
| Developer | Alibaba | StepFun |
| Released | 26 Aug 2026 | 18 Sep 2026 |
| Context window (tokens) | 256,000 | 1,000,000 |
| Compared at | Qwen3.8-Flash-Next | Step 5 Preview |
| Open weights | Yes | Yes |
| Price in / out / cached input per 1M | $0.150 / $0.470 / – | $1.00 / $2.70 / $0.105 |
Category by category
Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.
| Category | Qwen3.8-Flash-Next | Step 5 Preview | Leader |
|---|---|---|---|
| Reasoning | 44.5 | 52.7 | Step 5 Preview |
| Coding | 65.7 | 50.2 | Qwen3.8-Flash-Next |
| Agents & tools | 64.1 | 50.6 | Qwen3.8-Flash-Next |
| Maths | – | 59.7 | not comparable |
| Knowledge | 38.4 | 55.2 | Step 5 Preview |
| Human preference | – | 63.2 | not comparable |
| Multimodal | 62.2 | 59.4 | Qwen3.8-Flash-Next |
| Long context | 64.6 | 69.2 | Step 5 Preview |
| Composite | – | 40.8 | not comparable |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Maths, Instruction following, Human preference.
ModelIndex
Where sources disagree
Across the 18 benchmarks both have, Qwen3.8-Flash-Next leads 8, Step 5 Preview leads 10.
| Overall index | Qwen3.8-Flash-Next | Step 5 Preview | Leader |
|---|---|---|---|
| BenchLeader Index | 54.6 | 54.9 | neck and neck |
Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).
Price, speed and latency
List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | Qwen3.8-Flash-Next | Step 5 Preview |
|---|---|---|
| Blended price per 1M tokens | $0.230 | $1.43 |
| Input / output per 1M | $0.150 / $0.470 | $1.00 / $2.70 |
| Cached input per 1M | – | $0.105 |
| Price per task: chat reply | $0.0002 | $0.0012 |
| Price per task: summarize a 30-page report | $0.0021 | $0.014 |
| Price per task: code edit | $0.0016 | $0.010 |
| Price per task: agentic coding session | $0.011 | $0.071 |
| Price per task: structured extraction | $0.0004 | $0.0025 |
| Agentic coding session, input mostly cached | – | $0.030 |
| Cost per benchmark task (AA Index) | – | $2.35 |
| Output tokens per task (AA Index) | – | 113k |
| Of which reasoning | – | 80k |
| Time per task (AA Index) | – | 14 min |
| Output speed | 64 tok/s | 132 tok/s |
| Time to first answer | 125.1 s | 25.9 s |
| Time to a complete answer | 286.8 s | 57.3 s |
Each model is compared at its best-ranked reasoning-effort setting. Qwen3.8-Flash-Next scores 54.6 and takes 125 s to its first answer token. Step 5 Preview scores 54.9 and takes 26 s to its first answer token. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.
Every benchmark
88 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.
| Metric | Publisher | Qwen3.8-Flash-Next | Step 5 Preview |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 54.6 | 54.9 |
| Agents & tools score | BenchLeader | 64.1 | 50.6 |
| Coding score | BenchLeader | 65.7 | 50.2 |
| Knowledge score | BenchLeader | 38.4 | 55.2 |
| Long context score | BenchLeader | 64.6 | 69.2 |
| Multimodal score | BenchLeader | 62.2 | 59.4 |
| Reasoning score | BenchLeader | 44.5 | 52.7 |
| Composite score | BenchLeader | – | 40.8 |
| Human preference score | BenchLeader | – | 63.2 |
| Maths score | BenchLeader | – | 59.7 |
| Blended price $/M | BenchLeader | $0.230 | $1.43 |
| Output speed | BenchLeader | 64 tok/s | 132 tok/s |
| Time to first answer | BenchLeader | 125.1 s | 25.9 s |
| Context window | BenchLeader | 256,000 tokens | 1,000,000 tokens |
| Reasoning | |||
| Arena Hard Prompts | Arena (LMArena) | – | 1477 |
| LiveBench 2.0 Reasoning | LiveBench | – | 83.8% |
| LiveBench 1.0 Reasoning | LiveBench | 87.4% | – |
| GPQA Diamond (AA) | Artificial Analysis | 92.3% | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 38.0% | 46.5% |
| HLE-Diamond (AA) | Artificial Analysis | 16.3% | 24.3% |
| CritPt | Artificial Analysis | 11.1% | 20.9% |
| CritPt v2 hard subset (AA) | Artificial Analysis | – | 32.5% |
| MysteryMechanism | Vals AI | – | 18.5% |
| Coding | |||
| SciCode | SciCode | – | 58.9% |
| Arena Coding | Arena (LMArena) | – | 1506 |
| Arena WebDev | Arena (LMArena) | 1628 | 1571† |
| LiveBench 2.0 Coding | LiveBench | – | 85.6% |
| LiveBench 1.0 Coding | LiveBench | 72.5% | – |
| SciCode (AA) | Artificial Analysis | 50.6% | 58.9% |
| IOI | Vals AI | – | 38.1% |
| Code Migration | Vals AI | – | 39.0% |
| Code Migration v1.1 | Vals AI | – | 47.0% |
| SWE-bench Multimodal | Vals AI | – | 40.8% |
| Vibe Code Bench v1.1 | Vals AI | – | 70.4% |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 25.3% | 33.3% |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | 86.1% | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | – | 38.4% |
| FrontierCode 2.0 (AA) | Artificial Analysis | – | 12.0% |
| Agents & tools | |||
| Arena Agent | Arena (LMArena) | -4.4 | -2.9 |
Show all 88 benchmarks
| Metric | Publisher | Qwen3.8-Flash-Next | Step 5 Preview |
|---|---|---|---|
| Agents & tools | |||
| LiveBench Long Agentic | LiveBench | – | 19.4% |
| LiveBench Browser | LiveBench | – | 24.8% |
| LiveBench Agentic Coding | LiveBench | 61.6% | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 56.4% | 54.3% |
| GDPval-AA v2.1 | Artificial Analysis | 1636 | 1594 |
| τ³-Banking (AA) | Artificial Analysis | 45.4% | – |
| ITBench SRE (AA) | Artificial Analysis | – | 55.6% |
| Analyst Agent (AA) | Artificial Analysis | – | 35.0% |
| APEX-Agents (AA) | Artificial Analysis | – | 38.0% |
| Finance Agent v2 | Vals AI | – | 50.7% |
| Finance Agent v2.1 | Vals AI | – | 58.9% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | – | 12.9% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 9.2% |
| Legal Research Bench | Vals AI | – | 40.4% |
| SREBench | Vals AI | – | 0.4% |
| Tax Agent Bench | Vals AI | – | 27.3% |
| Terminal-Bench 4.0 (Vals) | Vals AI | – | 31.8% |
| Terminal-Bench Science | Vals AI | – | 2.9% |
| AutomationBench | Artificial Analysis | 55.9% | 51.0% |
| GDP.pdf (AA) | Artificial Analysis | 15.6% | 14.8% |
| EnterpriseOps-Gym | Artificial Analysis | – | 47.2% |
| AA-Briefcase v1.1 | Artificial Analysis | 1574 | 1425 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | – | 2.4% |
| CWE-Bench (AA) | Artificial Analysis | – | 52.5% |
| DeepSec-Bench (AA) | Artificial Analysis | – | 15.8% |
| CyberGym end-to-end (AA) | Artificial Analysis | – | 74.8% |
| Maths | |||
| ProofBench | Vals AI | – | 42.0% |
| LiveBench Mathematics | LiveBench | 85.8% | – |
| Arena Maths | Arena (LMArena) | – | 1466 |
| Knowledge | |||
| LiveBench 2.0 Data Analysis | LiveBench | – | 63.5% |
| LiveBench Research | LiveBench | – | 72.8% |
| LiveBench 1.0 Data Analysis | LiveBench | 74.2% | – |
| AA-Omniscience v1.1 | Artificial Analysis | -12.2 | 16.4 |
| BioMysteryBench | Vals AI | – | 67.4% |
| Excel Modeling Benchmark | Vals AI | – | 57.1% |
| Excel Modeling Benchmark v1.1 | Vals AI | – | 69.5% |
| Public Benefits Bench v1.1 | Vals AI | – | 58.1% |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 23.5% | 36.4% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 53.2% | 68.7% |
| Instruction following | |||
| LiveBench Language | LiveBench | 74.6% | – |
| LiveBench Instruction Following | LiveBench | 77.1% | – |
| Arena Instruction Following | Arena (LMArena) | – | 1446 |
| Human preference | |||
| Arena Text | Arena (LMArena) | – | 1448 |
| Arena Creative Writing | Arena (LMArena) | – | 1411 |
| Arena Multi-turn | Arena (LMArena) | – | 1442 |
| Arena Longer Queries | Arena (LMArena) | – | 1460 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | – | 1265 |
| LiveBench Design | LiveBench | – | 76.9% |
| MMMU-Pro | Artificial Analysis | 79.8% | 76.4% |
| Code2Video-AA | Artificial Analysis | – | 821 |
| AA-3D | Artificial Analysis | – | 854 |
| Long context | |||
| AA-LCR v1.1 | Artificial Analysis | 79.7% | 88.3% |
| MLCR | Artificial Analysis | 20.6% | 18.3% |
| Composite | |||
| LiveBench 2.0 | LiveBench | – | 58.6% |
| LiveBench 1.0 | LiveBench | 76.2% | – |
| AA Intelligence Index v5.0 | Artificial Analysis | – | 28.5 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 39.8 | 43.7 |
| AA Progress Index v1 | Artificial Analysis | 320.2 | 318.2 |
| Vals Index | Vals AI | – | 54.8 |
| Vals Finance Index | Vals AI | – | 55.4% |
| Vals Science Index | Vals AI | – | 29.6% |
| AA Cyber Index | Artificial Analysis | – | 47.7 |
| Metadata | |||
| AA Openness Index | Artificial Analysis | – | 33.3 |
Data as of 2026-11-06. Best configuration of each model; every score links to its source on the model pages.
What changed
- Step 5 Preview: 2 new results (Vals Index, Harvey's Legal Agent Benchmark v1.1).
- Qwen3.8-Flash-Next: 1 new result (AA-Briefcase v1.1); 1 score revision.
- Step 5 Preview: 1 score revision.
- Qwen3.8-Flash-Next: BenchLeader Index 58.7 → 54.7.
- Step 5 Preview: BenchLeader Index 58.0 → 54.7.
- Qwen3.8-Flash-Next: 2 new results (AA Progress Index v1, HLE-Diamond (AA)); 3 score revisions.
- Step 5 Preview: 5 new results (AA Intelligence Index v5.0, AA Progress Index v1, HLE-Diamond (AA)…); 2 score revisions; Step 5 Preview cost per task moved from $1.03/M to $2.35/M (128% up).
- Step 5 Preview: 3 score revisions.
- Step 5 Preview: 1 score revision.
- Step 5 Preview: 7 new results (Code Migration v1.1, Excel Modeling Benchmark v1.1, Finance Agent v2.1…).
Related comparisons
- Qwen3.8-Flash-Next vs GPT-6 Astra
- Qwen3.8-Flash-Next vs GPT-6.1 Sol
- Qwen3.8-Flash-Next vs Gemini 4 Argon
- Step 5 Preview vs Gemini 4 Argon
- Step 5 Preview vs Gemini 3.8 Flash
- Step 5 Preview vs Muse Spark 1.3
- Step 5 Preview vs DeepSeek V4.1 Flash 1017
- Qwen3.8-Flash-Next vs GPT-5.5
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Qwen3.8-Flash-Next vs Step 5 Preview: questions
- Is Qwen3.8-Flash-Next better than Step 5 Preview?
- Qwen3.8-Flash-Next and Step 5 Preview are neck and neck on quality (54.6 vs 54.9, within the ±10.4 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 6 Nov 2026, so check the category scores for your use.
- Qwen3.8-Flash-Next vs Step 5 Preview: which is cheaper, and will that change?
- Qwen3.8-Flash-Next is cheaper today: $0.230 against $1.43 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Qwen3.8-Flash-Next or Step 5 Preview better for coding?
- Qwen3.8-Flash-Next scores higher in coding (66 vs 50 on the category index, where 50 is average).
- Which is faster, Qwen3.8-Flash-Next or Step 5 Preview?
- Step 5 Preview streams 3×+ faster (132 vs 64 output tokens per second, Artificial Analysis 7-day median).
- Is Qwen3.8-Flash-Next or Step 5 Preview better for agents and tool use?
- Qwen3.8-Flash-Next scores higher in agentic tasks (64 vs 51 on the category index, where 50 is average). On Arena Agent, the top shared agents & tools benchmark, it is -4.4 for Qwen3.8-Flash-Next against -2.9 for Step 5 Preview.
More questions (4)
- Which is better for long documents?
- Step 5 Preview scores higher on long-context benchmarks (69 vs 65). Step 5 Preview accepts more context: 1,000,000 against 256,000 tokens for Qwen3.8-Flash-Next.
- Which has lower latency?
- Step 5 Preview answers first: 25.9 s against 125.1 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
- Do independent benchmarks agree on Qwen3.8-Flash-Next vs Step 5 Preview?
- Across the 18 benchmarks both have, Qwen3.8-Flash-Next leads 8, Step 5 Preview leads 10.
- When was Step 5 Preview released?
- Step 5 Preview was released on 18 Sep 2026.