together:inkling-small

8/8/2026, 8:54:31 AM · 15 tests · run 5364a0b2-e9f5-4597-8c67-9d94cf66715d

basic-generation

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

Produced 98 words (0 output tokens).

Result

{
  "wordCount": 98,
  "outputTokens": 0
}

9ab20129-318b-4f2c-91c3-edac1ada3e91

Aug 8, 2026, 08:54:06

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

992ms

TTFT

376ms

Tok/s

0.0

Tok/s (adj)

0.0

usage-cost

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

in=0 out=0 rt=— cost=$0.000000 (paid).

Result

{
  "inputTokens": 0,
  "outputTokens": 0,
  "reasoningTokens": null,
  "costUsd": 0,
  "paid": true
}

8f1ea2e2-c817-413d-9dd2-686139fe27a1

Aug 8, 2026, 08:54:07

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

803ms

TTFT

774ms

Tok/s

0.0

Tok/s (adj)

0.0

streaming-structure

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

mode=streaming, 132 stream events, 1 text block(s), 35 chars.

Result

{
  "mode": "streaming",
  "streamEventCount": 132,
  "textBlocks": 1,
  "textLen": 35
}

297c76ce-8cbd-42ce-ba98-dd4fab809668

Aug 8, 2026, 08:54:08

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

811ms

TTFT

663ms

Tok/s

0.0

Tok/s (adj)

0.0

complete-stream-parity

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

stream="Paris" · complete="Paris"

Result

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

5210d1de-e9bf-4dd9-b1b6-d5d47a07cb52

Aug 8, 2026, 08:54:08

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

276ms

TTFT

244ms

Tok/s

0.0

Tok/s (adj)

0.0

17c91def-7afc-4c22-bd1d-d578017f1776

Aug 8, 2026, 08:54:08

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000018

Mode

non-streaming

Cost breakdown

Input $0.000013Output $0.000005

Latency

144ms

Tok/s

27.8

reasoning-conformance

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

Class (from reasoning-config): toggleable.
off: out=0 · mid: out=0 (rt=—).

Result

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

58af5110-0133-44df-8990-fc712bed1b32

Aug 8, 2026, 08:54:10

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

2.0s

TTFT

168ms

Tok/s

0.0

Tok/s (adj)

0.0

fa12366b-06a9-4fdb-99d9-0dc929edb3b1

Aug 8, 2026, 08:54:16

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

5.2s

TTFT

2.5s

Tok/s

0.0

Tok/s (adj)

0.0

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
}

b5ce2779-0209-4660-ac0b-1829e9b39cf1

Aug 8, 2026, 08:54:16

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

327ms

Tok/s

0.0

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 of23 and17 is391."

Result

{
  "toolName": "multiply",
  "args": {
    "a": 23,
    "b": 17
  },
  "product": 391,
  "continuation": "The product of23 and17 is391.",
  "usesResult": true,
  "forced": true
}

0d17cedd-7408-44f0-bb70-c79045910559

Aug 8, 2026, 08:54:16

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

458ms

Tok/s

0.0

efe4dbb7-4ee7-4435-bc50-8b321edc3cfe

Aug 8, 2026, 08:54:17

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

293ms

TTFT

233ms

Tok/s

0.0

Tok/s (adj)

0.0

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.
reasoned=false, none token —; follow-up → "Balance for **acme-42**: **$1,234** Calculation (5% annual compound,7"

Result

{
  "signatureType": "none",
  "tokenCaptured": false,
  "reasoned": false,
  "answered": true,
  "continuation": "Balance for **acme-42**: **$1,234**\n\nCalculation (5% annual compound,7 years):\n\n\\[\n\\text{Future Value} =1234 \\times (1.05)^7\\]\n\n\\[\n(1.05)^7 "
}

1fa1b097-3473-4346-a7bd-c8fcaaf4ea25

Aug 8, 2026, 08:54:18

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

353ms

Tok/s

0.0

4ec5bc77-382c-4527-8aad-c96a32320d9b

Aug 8, 2026, 08:54:20

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000000

Mode

streaming

Latency

2.7s

TTFT

2.0s

Tok/s

0.0

Tok/s (adj)

0.0

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
}

b9059879-a619-42db-87f8-9f4100283e33

Aug 8, 2026, 08:54:21

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

273ms

TTFT

168ms

Tok/s

0.0

Tok/s (adj)

0.0

summarize-memex

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

Read ~0 input tokens, wrote 1147 words (0 output tokens) in 9113ms.

Result

{
  "wordCount": 1147,
  "inputTokens": 0,
  "outputTokens": 0,
  "latencyMs": 9113
}

f0514189-0acc-4a6d-a70d-917c8ca8cfda

Aug 8, 2026, 08:54:30

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

9.1s

TTFT

384ms

Tok/s

0.0

Tok/s (adj)

0.0

truncation

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

finishReason=length, out=0 tokens.

Result

{
  "finishReason": "length",
  "outputTokens": 0
}

fcb7454c-ca20-402b-83e4-99a3c4175fdf

Aug 8, 2026, 08:54:31

togetherai:thinkingmachines/Inkling-Small

Input

Output

Cost

$0.000000

Mode

streaming

Latency

944ms

TTFT

429ms

Tok/s

0.0

Tok/s (adj)

0.0

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"
}