GPT-5.4 mini vs Qwen3.6-27B

Updated October 19, 2026. Rank says GPT-5.4 mini is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.

Decision reading

GPT-5.4 mini has the higher public point estimate, 53.4 versus 48.57. Their conditional score ranges overlap. These ranges do not establish rank confidence. 7 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Model A

GPT-5.4 mini

OpenAI

53.4/100

Supported · Public rank #78

90% interval 42.6–64.2

Model B

Qwen3.6-27B

Alibaba

48.57/100

Estimated · Public rank #104

Conditional range 34.2–62.9

Shared results: 7
GPT-5.4 mini only: 15
Qwen3.6-27B only: 33
Like-for-like categories: 2 / 8

Supported: GPT-5.4 mini · Estimated: Qwen3.6-27B. Conditional ranges do not establish rank confidence.How the comparison works

Which one for your work

Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.

Agentic work
Tool use, computer use, and multi-step task completion
GPT-5.4 mini
GPT-5.4 mini has the higher public agentic point estimate, 35.9 to 29.1, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Confidence: limited
Long documents
Prompts that approach the documented context limit
GPT-5.4 mini
GPT-5.4 mini has the larger documented context window.
Confidence: documented

Show secondary and unsupported calls

Coding work
Code generation, repair, and software-engineering tasks
No clear pick
The like-for-like coding result is a practical tie on the public lane (within 0.5 points).
Confidence: limited
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates

Which one for a specific job

Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.

Pick the job you would use them for

The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.

35.1GPT-5.4 mini34.8Qwen3.6-27B

Like-for-like · BenchAlign v5.8

GPT-5.4 mini has the higher coding point estimate. Conditional score ranges do not establish rank confidence.

Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.

Open the coding leaderboard Shortlist models for this job

Same basis rules as the category table below

Shape of the matched evidence

Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.

Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.

HLEKnowledge
Normalized gap 17.5
GPT-5.4 mini:41.5%
Qwen3.6-27B:24%
GPT-5.4 mini source Qwen3.6-27B source
FrontierMath v2 (Tiers 1-3)Math
Normalized gap 16.1
GPT-5.4 mini:51.228%
Qwen3.6-27B:35.088%
Shared source
MMMU-ProMultimodal
Normalized gap 0.8
GPT-5.4 mini:76.6%
Qwen3.6-27B:75.8%
GPT-5.4 mini source Qwen3.6-27B source
Terminal-Bench 2.0Agentic
Normalized gap 0.7
GPT-5.4 mini:60%
Qwen3.6-27B:59.3%
GPT-5.4 mini source Qwen3.6-27B source
GPQAKnowledge
Normalized gap 0.2
GPT-5.4 mini:88%
Qwen3.6-27B:87.8%
GPT-5.4 mini source Qwen3.6-27B source

Bars run 0–100 on each benchmark’s normalized display scale

Category results, on a stated basis

Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.

Agentic

Like-for-like

GPT-5.4 mini: 35.9; Supported · #80/130
Qwen3.6-27B: 29.1; Supported · #97/130
Basis: BenchAlign v5.8 lane · 6 vs 6 public rows
Reading: GPT-5.4 mini leads · intervals overlap

Coding

Like-for-like

GPT-5.4 mini: 35.1; Supported · #78/151
Qwen3.6-27B: 34.8; Supported · #80/151
Basis: BenchAlign v5.8 lane · 5 vs 7 public rows
Reading: Practical tie

Multimodal

Directional only

GPT-5.4 mini: 73.8; #28/55
Qwen3.6-27B: 67.3; #39/55
Basis: Provisional lane · 1 vs 2 weighted rows
Reading: Directional only

Knowledge

Directional only

GPT-5.4 mini: 51.5; Supported · #77/186
Qwen3.6-27B: 46.8; Estimated · #92/186
Basis: BenchAlign v5.8 lane · 5 vs 6 public rows
Reading: Directional only

Instruction following

Directional only

GPT-5.4 mini: 89.9; #27/148
Qwen3.6-27B: 82.4; #55/148
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Directional only

Reasoning

Not comparable

GPT-5.4 mini: 43.7; Unranked · 4 rankable rows
Qwen3.6-27B: 78.1; Unranked · 2 rankable rows
Basis: Provisional lane · 1 vs 0 weighted rows
Reading: Not comparable

Multilingual

Not comparable

GPT-5.4 mini: Not ranked
Qwen3.6-27B: Not ranked
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Not comparable

Math

Not comparable

GPT-5.4 mini: 55.9; Unranked · 2 rankable rows
Qwen3.6-27B: 65.9; #8/12
Basis: Provisional lane · 2 vs 3 weighted rows
Reading: Not comparable

Category scores with the server-provided evidence basis for GPT-5.4 mini and Qwen3.6-27B
Category	GPT-5.4 mini	Qwen3.6-27B	Basis	Reading
Agentic	35.9Supported · #80/130	29.1Supported · #97/130	Like-for-likeBenchAlign v5.8 lane · 6 vs 6 public rows	GPT-5.4 mini leads · intervals overlap
Coding	35.1Supported · #78/151	34.8Supported · #80/151	Like-for-likeBenchAlign v5.8 lane · 5 vs 7 public rows	Practical tie
Multimodal	73.8#28/55	67.3#39/55	Directional onlyProvisional lane · 1 vs 2 weighted rows	Directional only
Knowledge	51.5Supported · #77/186	46.8Estimated · #92/186	Directional onlyBenchAlign v5.8 lane · 5 vs 6 public rows	Directional only
Instruction following	89.9#27/148	82.4#55/148	Directional onlyProvisional lane · 0 vs 0 weighted rows	Directional only
Reasoning	43.7Unranked · 4 rankable rows	78.1Unranked · 2 rankable rows	Not comparableProvisional lane · 1 vs 0 weighted rows	Not comparable
Multilingual	Not ranked	Not ranked	Not comparableProvisional lane · 0 vs 0 weighted rows	Not comparable
Math	55.9Unranked · 2 rankable rows	65.9#8/12	Not comparableProvisional lane · 2 vs 3 weighted rows	Not comparable

Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.

Supported evidence per lane · bars run 0–100Methodology

What each workload costs

Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.

Chat turn

1K fresh input + 500 output tokens

GPT-5.4 mini: $0.003; Fits in one request
Qwen3.6-27B: Self-hosted; infrastructure cost varies; Fits in one request

Qwen3.6-27B has no comparable published API token rate.

Repository review

50K fresh input + 3K output tokens

GPT-5.4 mini: $0.051; Fits in one request
Qwen3.6-27B: Self-hosted; infrastructure cost varies; Fits in one request

Qwen3.6-27B has no comparable published API token rate.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

GPT-5.4 mini: $0.075; Fits in one request
Qwen3.6-27B: Self-hosted; infrastructure cost varies; Fits in one request; Cached-input rate unavailable

Qwen3.6-27B has no comparable published API token rate.

Cached input falls back to the list input rate only where a cached rate is unpublished

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

SpecificationGPT-5.4 miniQwen3.6-27B

Context window

Maximum documented context; output-token limits may be lower.

GPT-5.4 mini

400K

OpenAI GPT-5.4 mini model documentation

Qwen3.6-27B

262K

API model ID

GPT-5.4 mini

gpt-5.4-mini

OpenAI GPT-5.4 mini model documentation

Qwen3.6-27B

Not sourced

Cached-input rate

A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.

GPT-5.4 mini

$0.075 per 1M cached input tokens

OpenAI pricing

Qwen3.6-27B

No comparable hosted API rate

Documented inputs

GPT-5.4 mini

text, image

OpenAI model catalog

Qwen3.6-27B

Not sourced

Documented outputs

GPT-5.4 mini

text

OpenAI model catalog

Qwen3.6-27B

Not sourced

Provider availability

GPT-5.4 mini

Generally Available · OpenAI Responses API

OpenAI model catalog

Qwen3.6-27B

Not sourced

Reasoning profile

GPT-5.4 mini

Reasoning

Qwen3.6-27B

Reasoning

Weight access

GPT-5.4 mini

Proprietary

Qwen3.6-27B

Open Weight

License

GPT-5.4 mini

Proprietary

Qwen3.6-27B

Open Weight

Release date

GPT-5.4 mini

2026-03-17

Qwen3.6-27B

2026-04-21

If you already use one of these models

Deployment change: The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
Quality signal: GPT-5.4 mini has the higher public point estimate, 53.4 versus 48.57. Their conditional score ranges overlap. These ranges do not establish rank confidence.
Workload cost: A complete comparable API-rate estimate is not available for both models.
Context tradeoff: GPT-5.4 mini has the larger documented window (400K).

Run the same representative tasks against both endpoints before changing production traffic.

Questions

Which is better, GPT-5.4 mini or Qwen3.6-27B?

GPT-5.4 mini has the higher public point estimate, 53.4 versus 48.57. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.

Which is better for coding, GPT-5.4 mini or Qwen3.6-27B?

The like-for-like coding row is a practical tie on the public lane, 35.1 against 34.8, inside the 0.5-point band BenchLM treats as level.

Which is better for agentic tasks, GPT-5.4 mini or Qwen3.6-27B?

GPT-5.4 mini has the higher public agentic tasks point estimate, 35.9 to 29.1, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which costs less, GPT-5.4 mini or Qwen3.6-27B?

Both models do not have comparable published API token rates, so this page does not name a universal price winner.

Which has the larger context window, GPT-5.4 mini or Qwen3.6-27B?

GPT-5.4 mini has the larger documented context window: 400K, compared with 262K.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

GPT-5.4 mini

API / mo$3,938

Self-host / moNot listed

Break-even—

Proprietary model — self-hosting not applicable.

Qwen3.6-27B

API / mo$0

Self-host / mo$298

Break-even—

Model the full break-even

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence55 rows

Agentic

Terminal-Bench 2.0
GPT-5.4 mini60%
Source
Qwen3.6-27B59.3%
Source
GPT-5.4 mini leads this result
OSWorld-Verified
GPT-5.4 mini72.1%
Source
Qwen3.6-27B—
Not directly comparable
MCP Atlas
GPT-5.4 mini57.7%
Source
Qwen3.6-27B—
Not directly comparable
Toolathlon
GPT-5.4 mini42.9%
Source
Qwen3.6-27B—
Not directly comparable
τ²-bench results
GPT-5.4 mini93.4%
Source
Qwen3.6-27B—
Not directly comparable
Terminal-Bench 2.1 (Vals)
GPT-5.4 mini54.7%
Source
Qwen3.6-27B—
Not directly comparable
Claw-Eval
GPT-5.4 mini—
Qwen3.6-27B72.4%
Source
Not directly comparable
QwenClawBench
GPT-5.4 mini—
Qwen3.6-27B53.4%
Source
Not directly comparable
QwenWebBench
GPT-5.4 mini—
Qwen3.6-27B1487
Source
Not directly comparable
AndroidWorld
GPT-5.4 mini—
Qwen3.6-27B70.3%
Source
Not directly comparable
Gert Labs
GPT-5.4 mini—
Qwen3.6-27B54.84%
Source
Not directly comparable

Coding

Terminal-Bench 2.0
GPT-5.4 mini60.0%
Source
Qwen3.6-27B59.3%
Source
GPT-5.4 mini leads this result
Vibe Code Bench
GPT-5.4 mini47.97%
Source
Qwen3.6-27B—
Not directly comparable
FrontierCode 1.1 Main
GPT-5.4 mini27.0%
Source
Qwen3.6-27B—
Not directly comparable
LiveCodeBench (Vals)
GPT-5.4 mini81.5%
Source
Qwen3.6-27B—
Not directly comparable
SWE-bench (Vals)
GPT-5.4 mini73.0%
Source
Qwen3.6-27B70.0%
Source
GPT-5.4 mini leads this result
SWE-bench Verified
GPT-5.4 mini—
Qwen3.6-27B77.2%
Source
Not directly comparable
SWE Multilingual
GPT-5.4 mini—
Qwen3.6-27B71.3%
Source
Not directly comparable
SWE-bench Pro
GPT-5.4 mini—
Qwen3.6-27B53.5%
Source
Not directly comparable
LiveCodeBench v6
GPT-5.4 mini—
Qwen3.6-27B83.9%
Source
Not directly comparable
NL2Repo
GPT-5.4 mini—
Qwen3.6-27B36.2%
Source
Not directly comparable

Reasoning

ARC-AGI-1
GPT-5.4 mini63.70%
Source
Qwen3.6-27B—
Not directly comparable
ARC-AGI-2
GPT-5.4 mini18.9%
Source
Qwen3.6-27B—
Not directly comparable

Multimodal

MMMU-Pro
GPT-5.4 mini76.6%
Source
Qwen3.6-27B75.8%
Source
GPT-5.4 mini leads this result
MMMU-Pro w/ Python
GPT-5.4 mini78%
Source
Qwen3.6-27B—
Not directly comparable
MMMU
GPT-5.4 mini—
Qwen3.6-27B82.9%
Source
Not directly comparable
RealWorldQA
GPT-5.4 mini—
Qwen3.6-27B84.1%
Source
Not directly comparable
DynaMath
GPT-5.4 mini—
Qwen3.6-27B85.6%
Source
Not directly comparable
MStar
GPT-5.4 mini—
Qwen3.6-27B81.4%
Source
Not directly comparable
SimpleVQA
GPT-5.4 mini—
Qwen3.6-27B56.1%
Source
Not directly comparable
CharXiv
GPT-5.4 mini—
Qwen3.6-27B78.4%
Source
Not directly comparable
CC-OCR
GPT-5.4 mini—
Qwen3.6-27B81.2%
Source
Not directly comparable
CountBench
GPT-5.4 mini—
Qwen3.6-27B97.8%
Source
Not directly comparable
RefCOCO (avg)
GPT-5.4 mini—
Qwen3.6-27B92.5%
Source
Not directly comparable
ERQA
GPT-5.4 mini—
Qwen3.6-27B62.5%
Source
Not directly comparable
Video-MME (with subtitle)
GPT-5.4 mini—
Qwen3.6-27B87.7%
Source
Not directly comparable
VideoMMMU
GPT-5.4 mini—
Qwen3.6-27B84.4%
Source
Not directly comparable
MLVU (M-Avg)
GPT-5.4 mini—
Qwen3.6-27B86.6%
Source
Not directly comparable
V*
GPT-5.4 mini—
Qwen3.6-27B94.7%
Source
Not directly comparable

Knowledge

GPQA
GPT-5.4 mini88%
Source
Qwen3.6-27B87.8%
Source
GPT-5.4 mini leads this result
HLE
GPT-5.4 mini41.5%
Source
Qwen3.6-27B24%
Source
GPT-5.4 mini leads this result
HLE w/o tools
GPT-5.4 mini28.2%
Source
Qwen3.6-27B—
Not directly comparable
GPQA Diamond (Vals)
GPT-5.4 mini83.1%
Source
Qwen3.6-27B—
Not directly comparable
MMLU-Pro (Vals)
GPT-5.4 mini84.6%
Source
Qwen3.6-27B—
Not directly comparable
MMLU-Pro
GPT-5.4 mini—
Qwen3.6-27B86.2%
Source
Not directly comparable
MMLU-Redux
GPT-5.4 mini—
Qwen3.6-27B93.5%
Source
Not directly comparable
SuperGPQA
GPT-5.4 mini—
Qwen3.6-27B66%
Source
Not directly comparable
C-Eval
GPT-5.4 mini—
Qwen3.6-27B91.4%
Source
Not directly comparable

Math

FrontierMath v2 (Tiers 1-3)
Shared source
GPT-5.4 mini51.228%
Qwen3.6-27B35.088%
GPT-5.4 mini leads this result
FrontierMath v2 (Tier 4)
GPT-5.4 mini9.756%
Source
Qwen3.6-27B—
Not directly comparable
HMMT Feb 2025
GPT-5.4 mini—
Qwen3.6-27B93.8%
Source
Not directly comparable
HMMT Nov 2025
GPT-5.4 mini—
Qwen3.6-27B90.7%
Source
Not directly comparable
HMMT Feb 2026
GPT-5.4 mini—
Qwen3.6-27B84.3%
Source
Not directly comparable
MMAnswerBench
GPT-5.4 mini—
Qwen3.6-27B80.8%
Source
Not directly comparable
AIME26
GPT-5.4 mini—
Qwen3.6-27B94.1%
Source
Not directly comparable

55 public results · 7 shared

Related comparisons

Compare API pricing Read the methodology Open the model selector

Watch GPT-5.4 mini vs Qwen3.6-27B

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 5,600+ readers.

Last updated October 19, 2026

GPT-5.4 mini vs Qwen3.6-27B

Decision reading

Which one for your work

Agentic work

Long documents

Coding work

Chat turn cost

Cache-heavy agent loop cost

Repository review cost

Which one for a specific job

What is actually comparable

Shape of the matched evidence

Category results, on a stated basis

Agentic

Coding

Multimodal

Knowledge

Instruction following

Reasoning

Multilingual

Math

What each workload costs

Chat turn

Repository review

Cache-heavy agent loop

Specification differences

Context window

API model ID

Cached-input rate

Documented inputs

Documented outputs

Provider availability

Reasoning profile

Weight access

License

Release date

If you already use one of these models

Questions

Which is better, GPT-5.4 mini or Qwen3.6-27B?

Which is better for coding, GPT-5.4 mini or Qwen3.6-27B?

Which is better for agentic tasks, GPT-5.4 mini or Qwen3.6-27B?

Which costs less, GPT-5.4 mini or Qwen3.6-27B?

Which has the larger context window, GPT-5.4 mini or Qwen3.6-27B?

Self-host vs API cost

Benchmark evidence

Agentic

Coding

Reasoning

Multimodal

Knowledge

Math

Related comparisons

Watch GPT-5.4 mini vs Qwen3.6-27B