Armand McQueen

basic-generation

ok

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

Produced 98 words (128 output tokens).

Result

{
  "wordCount": 98,
  "outputTokens": 128
}

98e21f8e-d8ae-468a-a700-147e4a55a3f2

Aug 18, 2026, 08:11:10

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

25

Output

128

Reasoning

0

Answer

128

Cost

$0.000862

Mode

streaming

Cost breakdown

Input $0.000063Output $0.000800

Latency

1.9s

TTFT

628ms

Tok/s

65.7

Tok/s (adj)

97.0

usage-cost

ok

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

in=60 out=65 rt=0 cost=$0.000556 (paid).

Result

{
  "inputTokens": 60,
  "outputTokens": 65,
  "reasoningTokens": 0,
  "costUsd": 0.00055625,
  "paid": true
}

d7e55ccf-495d-48d7-9e93-236fbc4989c0

Aug 18, 2026, 08:11:12

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

60

Output

65

Reasoning

0

Answer

65

Cost

$0.000556

Mode

streaming

Cost breakdown

Input $0.000150Output $0.000406

Latency

1.3s

TTFT

1.2s

Tok/s

51.1

Tok/s (adj)

1277.8

streaming-structure

ok

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

mode=streaming, 29 stream events, 1 text block(s), 35 chars.

Result

{
  "mode": "streaming",
  "streamEventCount": 29,
  "textBlocks": 1,
  "textLen": 35
}

9606e489-6dcb-414e-ba43-25c55277f864

Aug 18, 2026, 08:11:13

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

66

Output

78

Reasoning

0

Answer

78

Cost

$0.000652

Mode

streaming

Cost breakdown

Input $0.000165Output $0.000487

Latency

881ms

TTFT

671ms

Tok/s

88.5

Tok/s (adj)

372.1

complete-stream-parity

ok

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

stream="Paris" · complete="Paris"

Result

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

a9689926-d791-4f82-a414-707bfc84609b

Aug 18, 2026, 08:11:13

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

26

Output

2

Reasoning

0

Answer

2

Cost

$0.000077

Mode

streaming

Cost breakdown

Input $0.000065Output $0.000012

Latency

238ms

TTFT

213ms

Tok/s

8.4

Tok/s (adj)

79.3

47c51f53-a243-401d-9f7a-bc4733a40279

Aug 18, 2026, 08:11:13

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

26

Output

2

Reasoning

0

Answer

2

Cost

$0.000077

Mode

non-streaming

Cost breakdown

Input $0.000065Output $0.000012

Latency

372ms

Tok/s

5.4

reasoning-conformance

ok

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

Class (from reasoning-config): toggleable.
off: out=594 · mid: out=1050 (rt=0, rchars=1043).

Result

{
  "class": "toggleable",
  "mode": "deterministic",
  "offOutputTokens": 594,
  "midOutputTokens": 1050,
  "midReasoningTokens": 0
}

e73953bc-4b60-4e57-96d1-128687545024

Aug 18, 2026, 08:11:17

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

103

Output

594

Reasoning

0

Answer

594

Cost

$0.0040

Mode

streaming

Cost breakdown

Input $0.000257Output $0.0037

Latency

4.1s

TTFT

370ms

Tok/s

143.2

Tok/s (adj)

157.2

8afc9411-43dd-4387-baec-b2c242e684bb

Aug 18, 2026, 08:11:26

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

101

Output

1,050

Reasoning

0

Answer

1,050

Cost

$0.0068

Mode

streaming

Cost breakdown

Input $0.000253Output $0.0066

Latency

9.0s

TTFT

4.7s

Tok/s

116.9

Tok/s (adj)

243.5

single-tool-call

ok

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
}

6664b038-a907-41ff-95ce-d57f03f5eff7

Aug 18, 2026, 08:11:27

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

348

Output

38

Reasoning

0

Answer

38

Cost

$0.0011

Mode

streaming

Cost breakdown

Input $0.000870Output $0.000238

Latency

1.1s

Tok/s

34.6

tool-loop

ok

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
}

b9e65b3a-5fa6-4cf7-a0a8-d1d04e6616b8

Aug 18, 2026, 08:11:29

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

380

Output

86

Reasoning

0

Answer

86

Cost

$0.0015

Mode

streaming

Cost breakdown

Input $0.000950Output $0.000538

Latency

1.5s

Tok/s

59.3

8626202a-4382-4c71-ba51-1b8b410e12bd

Aug 18, 2026, 08:11:30

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

489

Output

41

Reasoning

0

Answer

41

Cost

$0.0015

Mode

streaming

Cost breakdown

Input $0.0012Output $0.000256

Latency

961ms

TTFT

834ms

Tok/s

42.7

Tok/s (adj)

324.1

web-search

ok

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

ok

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.
reasoned=false, none token —; follow-up → "Great! The balance for account 'acme-42' is **$1,234 USD**. ## Calcul"

Result

{
  "signatureType": "none",
  "tokenCaptured": false,
  "reasoned": false,
  "answered": true,
  "continuation": "Great! The balance for account 'acme-42' is **$1,234 USD**.\n\n## Calculation: 5% Annual Compound Growth Over 7 Years\n\nThe compound interest f"
}

c9869950-795a-4749-a7d1-1e6019b534d4

Aug 18, 2026, 08:11:33

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

368

Output

76

Reasoning

0

Answer

76

Cost

$0.0014

Mode

streaming

Cost breakdown

Input $0.000920Output $0.000475

Latency

1.1s

TTFT

865ms

Tok/s

68.8

Tok/s (adj)

318.1

93615145-9697-48d3-86a0-6fa877c8fda0

Aug 18, 2026, 08:11:38

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

471

Output

618

Reasoning

0

Answer

618

Cost

$0.0050

Mode

streaming

Cost breakdown

Input $0.0012Output $0.0039

Latency

5.3s

TTFT

439ms

Tok/s

116.5

Tok/s (adj)

127.0

prompt-cache

ok

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

ok

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
}

09c4bfab-4779-4f87-8f39-3f40abb939ec

Aug 18, 2026, 08:11:39

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

74

Output

23

Reasoning

0

Answer

23

Cost

$0.000329

Mode

streaming

Cost breakdown

Input $0.000185Output $0.000144

Latency

619ms

TTFT

514ms

Tok/s

37.2

Tok/s (adj)

218.6

summarize-memex

ok

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

Read ~10388 input tokens, wrote 534 words (709 output tokens) in 8339ms.

Result

{
  "wordCount": 534,
  "inputTokens": 10388,
  "outputTokens": 709,
  "latencyMs": 8339
}

df8f7e7f-ef6f-432e-beea-289a2681c9c1

Aug 18, 2026, 08:11:47

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

10,388

Output

709

Reasoning

0

Answer

709

Cost

$0.0304

Mode

streaming

Cost breakdown

Input $0.0260Output $0.0044

Latency

8.3s

TTFT

746ms

Tok/s

85.0

Tok/s (adj)

93.4

truncation

ok

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

finishReason=length, out=100 tokens.

Result

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

829e27c9-ae07-4079-a7d5-0973300e63e6

Aug 18, 2026, 08:11:48

togetherai:Qwen/Qwen3.8-2.4T-A95B

Input

31

Output

100

Reasoning

0

Answer

100

Cost

$0.000702

Mode

streaming

Cost breakdown

Input $0.000077Output $0.000625

Latency

1.3s

TTFT

269ms

Tok/s

75.6

Tok/s (adj)

95.0

streaming-only-rejection

ok

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

streamingOnly=false; complete() succeeded.

Result

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

abort

ok

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

pre-aborted signal → rejected (AbortError).

Result

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