Catalog Changes · 2026-10-14 00:05:32.354076

Qwen: Qwen3 235B A22B Thinking 2507 · 2026-10-14 00:05:32.354076 UTC

OpenRouter

6 Changed Fields

Field / Array Position	Previous Value	Observed Value
Benchmarks · Design Arena[0] · Elo	998	999
Benchmarks · Design Arena[1] · Elo	1037	1035
Benchmarks · Design Arena[2] · Elo	952	950
Benchmarks · Design Arena[2] · Rank	117	118
Benchmarks · Design Arena[4] · Elo	948	947
Benchmarks · Design Arena[4] · Rank	118	117

← Back To Results

Qwen: Qwen3 235B A22B Thinking 2507

qwen/qwen3-235b-a22b-thinking-2507

OpenRouter Catalog Changes

Last Observed: 2026-10-14 12:02:52.520334 UTC

Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...

Find A Field

Architecture And Modalities · 3

Architecture · Instruct Type API Details /architecture/instruct_type qwen3	qwen3
Architecture · Modality API Details /architecture/modality text->text	text->text
Architecture · Tokenizer API Details /architecture/tokenizer Qwen3	Qwen3

Benchmarks · 3

Benchmarks · Artificial Analysis · Agentic Index API Details /benchmarks/artificial_analysis/agentic_index 1.3	1.3
Benchmarks · Artificial Analysis · Coding Index API Details /benchmarks/artificial_analysis/coding_index 22.1	22.1
Benchmarks · Artificial Analysis · Intelligence Index API Details /benchmarks/artificial_analysis/intelligence_index 12.7	12.7

Catalog Identity And Metadata · 11

Canonical Slug API Details /canonical_slug qwen/qwen3-235b-a22b-thinking-2507	qwen/qwen3-235b-a22b-thinking-2507
Context Tokens API Details /context_length 128000	128,000
Created API Details /created 1753449557	2025-07-25 13:19:17 UTC
Description API Details /description Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...	Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...
Expiration Date API Details /expiration_date Explicit Null	Explicit Null
Hugging Face Id API Details /hugging_face_id Qwen/Qwen3-235B-A22B-Thinking-2507	Qwen/Qwen3-235B-A22B-Thinking-2507
Id API Details /id qwen/qwen3-235b-a22b-thinking-2507	qwen/qwen3-235b-a22b-thinking-2507
Knowledge Cutoff API Details /knowledge_cutoff 2025-06-30	2025-06-30
Links · Details API Details /links/details /api/v1/models/qwen/qwen3-235b-a22b-thinking-2507/endpoints	/api/v1/models/qwen/qwen3-235b-a22b-thinking-2507/endpoints
Name API Details /name Qwen: Qwen3 235B A22B Thinking 2507	Qwen: Qwen3 235B A22B Thinking 2507
Supported Voices API Details /supported_voices Explicit Null	Explicit Null

Default Parameters · 3

Default Parameters · Frequency Penalty API Details /default_parameters/frequency_penalty Explicit Null	Explicit Null
Default Parameters · Temperature API Details /default_parameters/temperature Explicit Null	Explicit Null
Default Parameters · Top P API Details /default_parameters/top_p Explicit Null	Explicit Null

Request Limits · 1

Per Request Limits API Details /per_request_limits Explicit Null	Explicit Null

Pricing · 2

Output Price · USD/Token API Details /pricing/completion 0.0000035	0.0000035
Input Price · USD/Token API Details /pricing/prompt 0.00000045	0.00000045

Reasoning · 1

Reasoning · Mandatory API Details /reasoning/mandatory True	True

Provider Limits · 3

Top Provider · Context Length API Details /top_provider/context_length 128000	128,000
Top Provider · Is Moderated API Details /top_provider/is_moderated False	False
Top Provider · Max Completion Tokens API Details /top_provider/max_completion_tokens 16384	16,384

API Structure And Container Types

Architecture API Details /architecture Object · See Fields	Object · See Fields
Architecture · Input Modalities API Details /architecture/input_modalities Array · See Items	Array · See Items
Architecture · Output Modalities API Details /architecture/output_modalities Array · See Items	Array · See Items
Benchmarks API Details /benchmarks Object · See Fields	Object · See Fields
Benchmarks · Artificial Analysis API Details /benchmarks/artificial_analysis Object · See Fields	Object · See Fields
Benchmarks · Design Arena API Details /benchmarks/design_arena Array · See Items	Array · See Items
Default Parameters API Details /default_parameters Object · See Fields	Object · See Fields
Links API Details /links Object · See Fields	Object · See Fields
Pricing API Details /pricing Object · See Fields	Object · See Fields
Reasoning API Details /reasoning Object · See Fields	Object · See Fields
Supported Parameters API Details /supported_parameters Array · See Items	Array · See Items
Top Provider API Details /top_provider Object · See Fields	Object · See Fields

Architecture · Input Modalities

1 items · source order preserved

text

API Details

Item	/architecture/input_modalities/*
1	text

Architecture · Output Modalities

1 items · source order preserved

text

API Details

Item	/architecture/output_modalities/*
1	text

Supported Parameters

11 items · source order preserved

frequency_penalty
include_reasoning
max_tokens
presence_penalty
reasoning
stop
temperature
tool_choice
tools
top_k
top_p

API Details

Item	/supported_parameters/*
1	frequency_penalty
2	include_reasoning
3	max_tokens
4	presence_penalty
5	reasoning
6	stop
7	temperature
8	tool_choice
9	tools
10	top_k
11	top_p

Benchmarks · Design Arena

6 items · source order preserved

Item	Arena	Category	Elo	Rank	Win Rate
1	models	3d	999	108	40.4
2	models	codecategories	1035	114	40.8
3	models	dataviz	950	118	32.6
4	models	gamedev	962	122	34.2
5	models	uicomponent	947	117	34
6	models	website	1055	114	42

API Details

Item	/benchmarks/design_arena/*/arena	/benchmarks/design_arena/*/category	/benchmarks/design_arena/*/elo	/benchmarks/design_arena/*/rank	/benchmarks/design_arena/*/win_rate
1	models	3d	999	108	40.4
2	models	codecategories	1035	114	40.8
3	models	dataviz	950	118	32.6
4	models	gamedev	962	122	34.2
5	models	uicomponent	947	117	34
6	models	website	1055	114	42