Skip to main content
AI Choice Engine

Model comparison

Claude Haiku 5.5 vs Qwen 3.8 27B

Anthropic against Qwen, compared on context, price, and attributed benchmark results.

Claude Haiku 5.5: profile checked October 25, 2026Qwen 3.8 27B: profile checked October 25, 2026Individual provider fields may changeEvidence confidence: Low — see receipts below

Anthropic

Claude Haiku 5.5

Budget

vs

Qwen

Qwen 3.8 27B

Budget · Open weights

AI model capability comparison
SpecificationClaude Haiku 5.5Qwen 3.8 27B
ProviderAnthropicQwen
TierBudgetBudget
Context windowWinner: 1M262K
Max output128KNot verifiedUnverified
Input / 1M tokensWinner: $0.10$0.50
Output / 1M tokensWinner: $0.50$3
WeightsNot verifiedUnverifiedOpen
ParametersNot verifiedUnverified27B dense VLM (Gated DeltaNet hybrid)
Reasoning levelsNot verifiedUnverifiedlow, medium, xhigh
Modalitiestext, imagetext, image, video
LicenseNot verifiedUnverifiedApache 2.0
API model idclaude-haiku-5-5qwen3.8-27b
ReleasedOctober 7, 2026Not verifiedUnverified
Artificial Analysis Intelligence Index [xhigh] · source date Qwen 3.8 27B: 2026-10-01Not verifiedUnverified34

Dates beside benchmark scores refer to the linked source’s publication or capture date unless that source explicitly states when it measured the result. A shared URL alone does not establish comparable runs; unverified setups are marked. Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.

Benchmark receipts

Where each score comes from, and how far it can be compared across models.

  • Source date2026-10-01
    Artificial Analysis (Qwen3.8 27B (xhigh))

    Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.

    Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.

Pricing tiers

Claude Haiku 5.5: Base rates apply to prompts up to 100,000 tokens. Above 100,000 tokens, input is $0.50 and output $2.50 per MTok. Cache reads are $0.01/$0.05; 5-minute writes $0.125/$0.625 and 1-hour writes $0.20/$1 for short/long prompts. Batch input/output receives a 50% discount. Standard output is 128K; 300K Batch output requires the documented beta header. Adaptive thinking defaults to medium effort.

Qwen 3.8 27B: Hosted qwen3.8-27b International lists USD/MTok input/output $0.50/$3; Beijing lists $0.424/$1.696. QwenCloud marketplace quotes implicit cache $0.10, explicit creation $0.625/read $0.05. Its hosted card lists 1M context and 131K maximum response; the Apache-2.0 downloadable checkpoint has 262,144 native context. August 19 is the hosted listing, not proof of exact weight-publication day. Recommended generation budgets do not establish a hard native output cap.

BudgetRecord checked October 25, 2026

Claude Haiku 5.5

Anthropic positions Haiku 5.5 for high-volume classification, extraction, routing and subagent tasks, with adaptive thinking and a 1M-token context window.

Best for

  • Classification and extraction
  • Routing
  • High-volume subagent tasks

Watch out

Prices increase fivefold above 100K prompt tokens. The newer tokenizer counts approximately 30% more tokens than Haiku 4.5 for the same text. Omit temperature, top_p and top_k. Thinking blocks are reusable only in the producing account or a linked account. Parameter count and weight availability remain unverified; no predecessor benchmark scores are inherited.

BudgetOpen weightsRecord checked October 25, 2026

Qwen 3.8 27B

Qwen 3.8 27B is Alibaba's compact deployment-friendly dense VLM — 262K native context (1M via YaRN), Apache 2.0 weights, distinct from the hosted Qwen 3.8 Max API.

Best for

  • Self-hosting on a single node
  • Multimodal input at small scale
  • Long-context on modest hardware

Watch out

Native thinking defaults xhigh and supports low/medium/xhigh or non-thinking; preserve_thinking defaults true. Exact checkpoint release day and hard native output cap remain unverified. Managed 1M limits do not transfer to local weights. The native card recommends 131,072 final output for complex work, not a guaranteed generation cap. Quantization, memory and context scaling need target-hardware validation; video is input, text is output.

When the cheaper one wins

Claude Haiku 5.5 is cheaper on output at $0.50 per million tokens against $3 for Qwen 3.8 27B — about 6.0×. Use the cheaper tier for classification, extraction, summarisation, and any task where the expensive model’s extra score does not change the accepted output. The expensive one only pays if your hardest task actually fails on the cheap tier. This compares standard short-context token rates, not total cost per accepted task. Include input, cache, context tiers, schedules, retries and review in your pilot.

Run the model picker

Evidence confidence: Low

How strong and complete the evidence behind this comparison is — not a prediction of which model is better.

  • Pricing verified on both sides — Input and output rates are verified for both models.
  • 1/5 core specs verified on both sides — Not published for at least one side: max output, release date, parameter count, reasoning levels.
  • No directly aligned benchmark setup — No benchmark has matching cited sources, source dates, units, effort and recorded run details on both models.
  • Verified within the last 90 days — Older of the two catalog checks was 1 day ago.
  • Source citations incomplete — At least one side lacks the two-source citation set.

Source receipts

Each catalog figure was checked against the provider or an independent second source on the date shown.

Related comparisons

Diving deeper on one model? Claude Haiku 5.5 · Qwen 3.8 27B

Common questions

Claude Haiku 5.5 vs Qwen 3.8 27B

Answered from the verified figures on this page rather than general guidance.

Is Claude Haiku 5.5 or Qwen 3.8 27B cheaper for input?
Claude Haiku 5.5 is cheaper at $0.10 per million input tokens, against $0.50 for Qwen 3.8 27B — roughly 5.0× the price. Estimate both input and output token volume for your workload. These are base rates; Claude Haiku 5.5: Base rates apply to prompts up to 100,000 tokens. Above 100,000 tokens, input is $0.50 and output $2.50 per MTok. Cache reads are $0.01/$0.05; 5-minute writes $0.125/$0.625 and 1-hour writes $0.20/$1 for short/long prompts. Batch input/output receives a 50% discount. Standard output is 128K; 300K Batch output requires the documented beta header. Adaptive thinking defaults to medium effort. Qwen 3.8 27B: Hosted qwen3.8-27b International lists USD/MTok input/output $0.50/$3; Beijing lists $0.424/$1.696. QwenCloud marketplace quotes implicit cache $0.10, explicit creation $0.625/read $0.05. Its hosted card lists 1M context and 131K maximum response; the Apache-2.0 downloadable checkpoint has 262,144 native context. August 19 is the hosted listing, not proof of exact weight-publication day. Recommended generation budgets do not establish a hard native output cap.
Is Claude Haiku 5.5 or Qwen 3.8 27B cheaper for output?
Claude Haiku 5.5 is cheaper at $0.50 per million output tokens, against $3 for Qwen 3.8 27B — roughly 6.0× the price. Estimate both input and output token volume for your workload. These are base rates; Claude Haiku 5.5: Base rates apply to prompts up to 100,000 tokens. Above 100,000 tokens, input is $0.50 and output $2.50 per MTok. Cache reads are $0.01/$0.05; 5-minute writes $0.125/$0.625 and 1-hour writes $0.20/$1 for short/long prompts. Batch input/output receives a 50% discount. Standard output is 128K; 300K Batch output requires the documented beta header. Adaptive thinking defaults to medium effort. Qwen 3.8 27B: Hosted qwen3.8-27b International lists USD/MTok input/output $0.50/$3; Beijing lists $0.424/$1.696. QwenCloud marketplace quotes implicit cache $0.10, explicit creation $0.625/read $0.05. Its hosted card lists 1M context and 131K maximum response; the Apache-2.0 downloadable checkpoint has 262,144 native context. August 19 is the hosted listing, not proof of exact weight-publication day. Recommended generation budgets do not establish a hard native output cap.
Which has the larger context window, Claude Haiku 5.5 or Qwen 3.8 27B?
Claude Haiku 5.5 accepts 1M tokens against 262K for Qwen 3.8 27B. This only matters if you routinely send very long documents or large codebases.
Should I use Claude Haiku 5.5 or Qwen 3.8 27B?
Both are classified in our budget tier; this editorial grouping does not imply equal capability. Claude Haiku 5.5 is a candidate for classification and extraction; Qwen 3.8 27B is a candidate for self-hosting on a single node. Compare accuracy, tool reliability, latency and cost on representative tasks.

Next step

Choosing between them

The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.

  • Input price: Claude Haiku 5.5 costs $0.10 per 1M tokens versus $0.50 for Qwen 3.8 27B — a 5x difference at the headline tier.
  • Context: Claude Haiku 5.5 takes 1M against 262K for Qwen 3.8 27B — only decisive if your prompts approach the smaller window.
Run the model pickerEstimate stack costAll model comparisons

Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.