deepinfra:qwen3.6-35b-a3b

8/17/2026, 7:50:57 AM · 15 tests · run 875c0fbd-f24d-459c-a4d4-9668719279e4

basic-generation

Plain streaming generation with a soft length constraint (~100 words).

Produced 98 words (121 output tokens).

Result

{
  "wordCount": 98,
  "outputTokens": 121
}

d89e0437-7043-474c-bb05-6f56994b62b4

Aug 17, 2026, 07:49:34

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

121

Reasoning

Answer

121

Cost

$0.000117

Mode

streaming

Cost breakdown

Input $0.000003Output $0.000115

Latency

2.3s

TTFT

309ms

Tok/s

53.5

Tok/s (adj)

61.9

usage-cost

Asserts usage tokens populate and cost computes from the ChatReq contract.

in=18 out=353 rt=0 cost=$0.000337 (paid).

Result

{
  "inputTokens": 18,
  "outputTokens": 353,
  "reasoningTokens": 0,
  "costUsd": 0.00033715,
  "paid": true
}

06d446ea-aef6-44bf-9a1f-3407ba5ebbdf

Aug 17, 2026, 07:49:39

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

353

Reasoning

Answer

353

Cost

$0.000337

Mode

streaming

Cost breakdown

Input $0.000002Output $0.000335

Latency

5.1s

TTFT

5.0s

Tok/s

69.7

Tok/s (adj)

6592.0

streaming-structure

Asserts a streaming call captures events that reconstruct into the final text.

mode=streaming, 309 stream events, 1 text block(s), 35 chars.

Result

{
  "mode": "streaming",
  "streamEventCount": 309,
  "textBlocks": 1,
  "textLen": 35
}

4005f07e-f678-403d-8660-70a3836ff74a

Aug 17, 2026, 07:49:44

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

308

Reasoning

Answer

308

Cost

$0.000295

Mode

streaming

Cost breakdown

Input $0.000002Output $0.000293

Latency

5.1s

TTFT

4.5s

Tok/s

60.5

Tok/s (adj)

562.8

complete-stream-parity

Same prompt via stream() and complete() both return valid, matching results.

stream="Paris" · complete="Paris"

Result

{
  "streamText": "Paris",
  "completeText": "Paris"
}

eabc69c4-f154-440f-aabf-5e46cbddfacd

Aug 17, 2026, 07:49:45

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

Reasoning

Answer

Cost

$0.000005

Mode

streaming

Cost breakdown

Input $0.000003Output $0.000002

Latency

343ms

TTFT

333ms

Tok/s

5.8

Tok/s (adj)

192.1

b6f2cfbc-9ccd-4b43-b409-3d4b0ad4b28e

Aug 17, 2026, 07:49:45

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

Reasoning

Answer

Cost

$0.000005

Mode

non-streaming

Cost breakdown

Input $0.000003Output $0.000002

Latency

552ms

Tok/s

3.6

reasoning-conformance

Asserts the model's reasoning knob behaves as its reasoning-config declares.

Class (from reasoning-config): toggleable.
off: out=924 · mid: out=2759 (rt=0, rchars=5309).

Result

{
  "class": "toggleable",
  "mode": "deterministic",
  "offOutputTokens": 924,
  "midOutputTokens": 2759,
  "midReasoningTokens": 0
}

079f13f7-c605-4112-a637-0f25ec19ca8d

Aug 17, 2026, 07:49:59

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

103

Output

924

Reasoning

Answer

924

Cost

$0.000888

Mode

streaming

Cost breakdown

Input $0.000010Output $0.000878

Latency

14.1s

TTFT

681ms

Tok/s

65.4

Tok/s (adj)

68.7

5ee3a24f-1dce-4365-a352-9d80004b8310

Aug 17, 2026, 07:50:34

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

101

Output

2,759

Reasoning

Answer

2,759

Cost

$0.0026

Mode

streaming

Cost breakdown

Input $0.000010Output $0.0026

Latency

34.7s

TTFT

27.6s

Tok/s

79.4

Tok/s (adj)

387.4

single-tool-call

Offers a calculator tool and checks the model calls it for an arithmetic question.

Called "multiply" with args {"a":23,"b":17}.

Result

{
  "toolCalled": true,
  "toolName": "multiply",
  "args": {
    "a": 23,
    "b": 17
  },
  "correct": true
}

aef4803e-412f-4890-bef4-1f8b11d60154

Aug 17, 2026, 07:50:35

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

348

Output

Reasoning

Answer

Cost

$0.000071

Mode

streaming

Cost breakdown

Input $0.000035Output $0.000036

Latency

880ms

Tok/s

43.2

tool-loop

Forces a tool call, feeds the result back, and checks the continuation uses it.

step1 (required): multiply({"a":23,"b":17}) → step2: " The product of 23 multiplied by 17 is 391."

Result

{
  "toolName": "multiply",
  "args": {
    "a": 23,
    "b": 17
  },
  "product": 391,
  "continuation": "\n\nThe product of 23 multiplied by 17 is 391.",
  "usesResult": true,
  "forced": true
}

ffb98da5-1595-4789-94bb-17d6683b46a9

Aug 17, 2026, 07:50:39

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

342

Output

387

Reasoning

Answer

387

Cost

$0.000402

Mode

streaming

Cost breakdown

Input $0.000034Output $0.000368

Latency

4.5s

Tok/s

86.0

67019071-7baf-4f37-a417-2a9829772a1d

Aug 17, 2026, 07:50:40

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

752

Output

Reasoning

Answer

Cost

$0.000115

Mode

streaming

Cost breakdown

Input $0.000075Output $0.000040

Latency

1.0s

TTFT

854ms

Tok/s

40.9

Tok/s (adj)

241.0

web-search

Enables server-side web search and checks invocation counts, the per-request cost surcharge, and provider-executed output blocks.

Skipped: model does not support the catalog's web_search tool version.

Result

{
  "skipped": "no server-side web search support"
}

reasoning-tool-continuity

Reasoning model: continuity token captured on a tool call and replayed so the follow-up succeeds.

Step 1 didn't engage reasoning (adaptive discretion) — signature not exercised this run.
follow-up succeeded but didn't echo the literal balance value.
reasoned=false, none token —; follow-up → ""

Result

{
  "signatureType": "none",
  "tokenCaptured": false,
  "reasoned": false,
  "answered": false,
  "continuation": ""
}

51ec5319-ab77-4175-9be5-1fbc8b7cc597

Aug 17, 2026, 07:50:47

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

368

Output

222

Reasoning

Answer

222

Cost

$0.000248

Mode

streaming

Cost breakdown

Input $0.000037Output $0.000211

Latency

2.7s

Tok/s

80.9

89b9a237-4b7f-445b-9db6-307e0360101f

Aug 17, 2026, 07:50:48

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

617

Output

Reasoning

Answer

Cost

$0.000090

Mode

streaming

Cost breakdown

Input $0.000062Output $0.000028

Latency

584ms

Tok/s

51.4

prompt-cache

Repeat a large prefix; assert cache-read tokens are reported and cost drops.

Skipped: provider has no (known) prompt-cache reporting.

Result

{
  "skipped": "cacheStyle=none"
}

json-output

Requests a strict JSON object and checks it parses with the required keys.

Output parsed as JSON.
All required keys present.
Output is bare JSON (no leading prose/fence).

Result

{
  "parsed": {
    "name": "Anthropic",
    "founded": 2021,
    "headquarters": "San Francisco, California"
  },
  "validJson": true,
  "missingKeys": [],
  "bareJson": true
}

088e0cc8-5bfb-4070-9b09-76a2f36ec389

Aug 17, 2026, 07:50:48

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

Reasoning

Answer

Cost

$0.000034

Mode

streaming

Cost breakdown

Input $0.000007Output $0.000027

Latency

414ms

TTFT

147ms

Tok/s

67.6

Tok/s (adj)

105.0

summarize-memex

Summarizes an ~11k-token essay to ~500 words (long input, reasoning minimized).

Read ~10388 input tokens, wrote 461 words (572 output tokens) in 6497ms.

Result

{
  "wordCount": 461,
  "inputTokens": 10388,
  "outputTokens": 572,
  "latencyMs": 6497
}

c7e8cee5-7bb0-4854-b56d-d71e234f89f2

Aug 17, 2026, 07:50:55

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

10,388

Output

572

Reasoning

Answer

572

Cost

$0.0016

Mode

streaming

Cost breakdown

Input $0.0010Output $0.000543

Latency

6.5s

TTFT

552ms

Tok/s

88.0

Tok/s (adj)

96.2

truncation

Tiny maxTokens on a long prompt yields finishReason:"length".

finishReason=length, out=100 tokens.

Result

{
  "finishReason": "length",
  "outputTokens": 100
}

e42fc892-6c64-4b32-8a59-bdf1118b1e77

Aug 17, 2026, 07:50:56

deepinfra:Qwen/Qwen3.6-35B-A3B

Input

Output

100

Reasoning

Answer

100

Cost

$0.000098

Mode

streaming

Cost breakdown

Input $0.000003Output $0.000095

Latency

1.6s

TTFT

403ms

Tok/s

62.4

Tok/s (adj)

83.4

streaming-only-rejection

Streaming-only models reject complete(); all others accept it.

streamingOnly=false; complete() succeeded.

Result

{
  "streamingOnly": false,
  "completeThrew": false
}

abort

A pre-aborted signal rejects the call (abort is threaded + honored).

pre-aborted signal → rejected (AbortError).

Result

{
  "rejected": true,
  "errorName": "AbortError"
}