Skip to content
Get an API key

Live · serverless · 3 regions

Qwen2.5 Coder 32B Instruct

by Alibaba

A code model for completion, refactoring and test generation across more than 40 programming languages.

  • code
  • FP8
  • dedicated
Serverless priceUSD
Input
$0.28per 1M tokens
Output
$0.44per 1M tokens

Billed per token, no minimums.

Reference

Specifications

Model ID
qwen2-5-coder-32b
Endpoint
POST /v1/chat/completions
Publisher
Alibaba
Tasks
code
Served context
128K tokens
Quantization
FP8
Decode speed
125 tok/smedian, one request
p50 first token
140 mslast 24 hours
Regions
us-east · us-west · eu-west
Dedicated
1× NVIDIA H100 per replica

Last 24 hours

Latency and speed

Median of each 15-minute step on the shared pool, measured at the API edge. Time to first token includes queueing; decode speed is for one request.

Time to first token · ms

140 msmedian

p95 176 ms

Decode speed · tok/s

125 tok/smedian

p5 109 tok/s

Show the data
Median of each 15-minute step over the last 24 hours, newest first
StepTime to first tokenDecode speed
15m ago147 ms119 tok/s
30m ago142 ms125 tok/s
45m ago141 ms125 tok/s
1h ago142 ms128 tok/s
1h 15m ago150 ms123 tok/s
1h 30m ago145 ms124 tok/s
1h 45m ago149 ms120 tok/s
2h ago143 ms125 tok/s
2h 15m ago146 ms121 tok/s
2h 30m ago206 ms94 tok/s
2h 45m ago185 ms101 tok/s
3h ago143 ms120 tok/s
3h 15m ago147 ms123 tok/s
3h 30m ago149 ms119 tok/s
3h 45m ago151 ms121 tok/s
4h ago152 ms119 tok/s
4h 15m ago140 ms123 tok/s
4h 30m ago140 ms127 tok/s
4h 45m ago144 ms120 tok/s
5h ago142 ms122 tok/s
5h 15m ago150 ms118 tok/s
5h 30m ago147 ms125 tok/s
5h 45m ago146 ms126 tok/s
6h ago139 ms122 tok/s
6h 15m ago147 ms123 tok/s
6h 30m ago138 ms127 tok/s
6h 45m ago142 ms123 tok/s
7h ago144 ms123 tok/s
7h 15m ago149 ms123 tok/s
7h 30m ago148 ms123 tok/s
7h 45m ago138 ms123 tok/s
8h ago141 ms122 tok/s
8h 15m ago140 ms124 tok/s
8h 30m ago141 ms121 tok/s
8h 45m ago143 ms127 tok/s
9h ago147 ms125 tok/s
9h 15m ago139 ms124 tok/s
9h 30m ago132 ms126 tok/s
9h 45m ago133 ms132 tok/s
10h ago139 ms129 tok/s
10h 15m ago136 ms131 tok/s
10h 30m ago132 ms128 tok/s
10h 45m ago187 ms106 tok/s
11h ago131 ms130 tok/s
11h 15m ago142 ms123 tok/s
11h 30m ago136 ms130 tok/s
11h 45m ago139 ms128 tok/s
12h ago131 ms131 tok/s
12h 15m ago130 ms130 tok/s
12h 30m ago132 ms130 tok/s
12h 45m ago141 ms124 tok/s
13h ago139 ms126 tok/s
13h 15m ago130 ms133 tok/s
13h 30m ago128 ms130 tok/s
13h 45m ago140 ms124 tok/s
14h ago139 ms129 tok/s
14h 15m ago128 ms131 tok/s
14h 30m ago132 ms128 tok/s
14h 45m ago137 ms128 tok/s
15h ago130 ms133 tok/s
15h 15m ago131 ms131 tok/s
15h 30m ago132 ms129 tok/s
15h 45m ago133 ms129 tok/s
16h ago129 ms128 tok/s
16h 15m ago133 ms129 tok/s
16h 30m ago138 ms125 tok/s
16h 45m ago130 ms131 tok/s
17h ago129 ms133 tok/s
17h 15m ago140 ms127 tok/s
17h 30m ago132 ms129 tok/s
17h 45m ago139 ms123 tok/s
18h ago176 ms110 tok/s
18h 15m ago132 ms130 tok/s
18h 30m ago136 ms130 tok/s
18h 45m ago139 ms123 tok/s
19h ago138 ms123 tok/s
19h 15m ago141 ms128 tok/s
19h 30m ago134 ms126 tok/s
19h 45m ago143 ms127 tok/s
20h ago132 ms129 tok/s
20h 15m ago132 ms133 tok/s
20h 30m ago173 ms109 tok/s
20h 45m ago133 ms132 tok/s
21h ago193 ms104 tok/s
21h 15m ago146 ms121 tok/s
21h 30m ago138 ms127 tok/s
21h 45m ago148 ms118 tok/s
22h ago140 ms129 tok/s
22h 15m ago141 ms121 tok/s
22h 30m ago140 ms124 tok/s
22h 45m ago142 ms126 tok/s
23h ago140 ms127 tok/s
23h 15m ago146 ms119 tok/s
23h 30m ago148 ms120 tok/s
23h 45m ago137 ms126 tok/s
24h ago138 ms127 tok/s

API

Make a request

The same request shape as every other model on /v1/chat/completions. Switching between them is a change to one string.

curl https://api.weir.example/v1/chat/completions \
  -H "Authorization: Bearer $WEIR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen2-5-coder-32b",
    "messages": [
      { "role": "user", "content": "What does a KV cache do?" }
    ]
  }'
POST /v1/chat/completionsapi.weir.example

Dedicated

Run it on GPUs reserved for you

Replica
1× H100
Replica rate
$2.49 / hour
Peak throughput
5,200 tok/s
Break-even, 1 replica
5.05B tok/mo

Replicas reserved for you, billed by the second, with a 99.95% monthly uptime SLA. Break-even is where one replica costs the same as serverless, at half input and half output tokens.

Run on dedicated GPUsCompare costs

Compare

Similar models