Unnamed variant

Registry metadata

variant_id
openrouter::qwen/qwen3-8b-04-28::structured
model_id
qwen/qwen3-8b
canonical_id
qwen/qwen3-8b-04-28
config_key
structured
variant_key
structured_output
interface
openrouter
first_seen_at
2026-03-13T04:16:17.881000+00:00
inference_config
[object Object]
label_suffix
(Structured Output)
variant_note
Response via response_format=json_schema; schema-validated JSON without tools
use_tools
false
use_reasoning
false
use_structured_output
true
use_web_search
false
use_low_temp
false
variant_last_seen_at
2026-05-25T07:43:32.395472+00:00
is_claimed_valid
true
is_retired
false
name
Qwen: Qwen3 8B
org
qwen
org_name
Alibaba DAMO Academy
country
China
city
Hangzhou
org_type
big-tech
open_weights
null
context_length
131072
max_completion_tokens
8192
tokenizer
Qwen3
pricing_input_per_1m
0.049999999999999996
pricing_output_per_1m
0.39999999999999997
tags
text_generation
release_date
null
expiration_date
null
param_count_b
8
active_param_count_b
null
is_moe
false
specialization
null
input_modalities
text
output_modalities
text
supported_parameters
frequency_penalty,include_reasoning,logit_bias,max_tokens,min_p,presence_penalty,reasoning,repetition_penalty,response_format,seed,stop,structured_outputs,temperature,tool_choice,tools,top_k,top_p
rate_limit_rpm
null
rate_limit_rpd
null
rate_limit_tpm
null
rate_limit_source
null
provenance_notes
Qwen series. Extremely strong Chinese coverage. Competitive English. Strong coding and math. Released many open-weight variants.
is_alias
true
source
openrouter
model_first_seen_at
2026-03-15T23:31:51.523000+00:00
model_last_seen_at
2026-05-25T07:43:32.395472+00:00
is_available
true
unavailable_reason
null
last_checked_at
2026-04-12T14:01:59.872642+00:00
last_latency_ms
10613
first_unavailable_at
2026-03-15T19:48:48.939000+00:00
arch_id
qwen3-dense
decoder_type
Dense
attention
GQA with QK-Norm
arch_highlight
Reference dense Qwen stack with QK-Norm and 8 KV heads.
tech_report_url
https://arxiv.org/pdf/2505.09388
hf_config_url
https://huggingface.co/Qwen/Qwen3-32B/blob/main/config.json
variant_status
available
reason_code
null
reason_detail
null
status_http_status
200
status_checked_at
2026-04-12T14:07:57.387432+00:00
status_source
live_traffic
claimed_capabilities
frequency_penalty,include_reasoning,logit_bias,max_tokens,min_p,presence_penalty,reasoning,repetition_penalty,response_format,seed,stop,structured_outputs,temperature,tool_choice,tools,top_k,top_p
required_capabilities
structured_outputs
verified_capabilities
structured_outputs
route_count
0

Similar variants