All times Eastern (ET)Provider pages checked daily at
ProSign inRSS

Migrating from qwen3.5-27b to qwen3.8-27b

What changes when you move from qwen3.5-27b to qwen3.8-27b: input price +66.7%, output price +25%, context window 262,144 → 1,000,000 tokens.

Status
Active
Shutdown
none announced
Time left
—
Status
Active
Shutdown
none announced
Time left
—

Prices

USD per 1M tokens, standard tier
Typeqwen3.5-27bqwen3.8-27bChange
Input$0.30first seen October 5, 2026 · source$0.50first seen October 5, 2026 · source+66.7%
Cached inputNot published$0.10first seen October 5, 2026 · source—
Output$2.40first seen October 5, 2026 · source$3.00first seen October 5, 2026 · source+25%

At 1M input and 250K output tokens a month: $0.90 on qwen3.5-27b, $1.25 on qwen3.8-27b (+38.9%).

Other tiers, prompt-size bands and modalities are on each model's page.

Limits

Tokens
Limitqwen3.5-27bqwen3.8-27bChange
Context window262,144 tokensin place when tracking began (October 6, 2026) · source1,000,000 tokensin place when tracking began (October 6, 2026) · source+281.5%
Max output tokens65,536 tokensin place when tracking began (October 6, 2026) · source131,072 tokensin place when tracking began (October 6, 2026) · source+100%

Benchmarks

Same benchmark, source and setting on both sides
Benchmarkqwen3.5-27bqwen3.8-27bChange
Arena TextArena. Ranges don't overlap.1409 ±41438 ±5+29
Arena Text: CodingArena. Ranges don't overlap.1452 ±71499 ±8+47
Arena WebDevArena. Ranges don't overlap.1358 ±81594 ±7+236
Arena VisionArena. Ranges don't overlap.1217 ±61244 ±8+27

A score from one source is never compared with another source's, and company-reported scores are compared only within the same company; settings named where they differ. Data: Arena (arena.ai), CC BY 4.0.

Features and parameters

Compare models
  • Prompt caching: not supported on qwen3.5-27b, supported on qwen3.8-27b.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source
  • Fine-tuning: supported on qwen3.5-27b, not supported on qwen3.8-27b.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source

Same on both

  • Structured outputs: supported on both.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source
  • Function calling: supported on both.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source
  • Batch API: not supported on both.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source
  • Image input: supported on both.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source
  • Built-in web search: supported on both.qwen3.5-27b: in place when tracking began (October 6, 2026) · source · qwen3.8-27b: in place when tracking began (October 6, 2026) · source

Not published for qwen3.5-27b: PDF input, Reasoning / thinking, temperature parameter, top_p parameter, logprobs.

Not published for qwen3.8-27b: PDF input, Reasoning / thinking, temperature parameter, top_p parameter, logprobs.

Aliases

All aliases

No aliases point at either model.

Checklist

  1. Change the model string from qwen3.5-27b to qwen3.8-27b.
  2. Re-check costs: qwen3.8-27b costs more than qwen3.5-27b (input +66.7%, output +25%).

Compared from AIWire's published data when this page loaded. Each fact links to the page it came from. Model pages: qwen3.5-27b, qwen3.8-27b.