together:inkling-small

8/8/2026, 9:07:08 AM · 15 tests · run 71360646-3765-4f66-912b-55854f3e04c8

basic-generation

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

Produced 121 words (156 output tokens).

Result

{
  "wordCount": 121,
  "outputTokens": 156
}

89ddaf96-0666-40db-a910-56f0ce225c0c

Aug 8, 2026, 09:06:45

togetherai:thinkingmachines/Inkling-Small

Input

Output

156

Reasoning

Answer

156

Cost

$0.000199

Mode

streaming

Cost breakdown

Input $0.000012Output $0.000187

Latency

1.3s

TTFT

504ms

Tok/s

118.0

Tok/s (adj)

190.8

usage-cost

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

in=23 out=132 rt=121 cost=$0.000170 (paid).

Result

{
  "inputTokens": 23,
  "outputTokens": 132,
  "reasoningTokens": 121,
  "costUsd": 0.0001699,
  "paid": true
}

10142f09-c4a4-4713-8216-73eeb0d5547b

Aug 8, 2026, 09:06:46

togetherai:thinkingmachines/Inkling-Small

Input

Output

132

Reasoning

121

Answer

Cost

$0.000170

Mode

streaming

Cost breakdown

Input $0.000012Output $0.000158

Latency

858ms

TTFT

828ms

Tok/s

153.8

Tok/s (adj)

371.2

streaming-structure

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

mode=streaming, 147 stream events, 1 text block(s), 35 chars.

Result

{
  "mode": "streaming",
  "streamEventCount": 147,
  "textBlocks": 1,
  "textLen": 35
}

bc4b826b-55dc-44ec-addc-41e5ad54fa28

Aug 8, 2026, 09:06:46

togetherai:thinkingmachines/Inkling-Small

Input

Output

145

Reasoning

110

Answer

Cost

$0.000188

Mode

streaming

Cost breakdown

Input $0.000014Output $0.000174

Latency

865ms

TTFT

717ms

Tok/s

167.6

Tok/s (adj)

236.2

complete-stream-parity

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

stream="Paris" · complete="Paris"

Result

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

cae4df79-3e7f-466d-8763-5e7fdc9e539e

Aug 8, 2026, 09:06:47

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000018

Mode

streaming

Cost breakdown

Input $0.000013Output $0.000005

Latency

178ms

TTFT

175ms

Tok/s

22.5

Tok/s (adj)

1212.1

68984977-c496-410a-ba14-41bf135430f3

Aug 8, 2026, 09:06:47

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000018

Mode

non-streaming

Cost breakdown

Input $0.000013Output $0.000005

Latency

135ms

Tok/s

29.6

reasoning-conformance

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

Class (from reasoning-config): toggleable.
off: out=375 · mid: out=1309 (rt=753).

Result

{
  "class": "toggleable",
  "mode": "deterministic",
  "offOutputTokens": 375,
  "midOutputTokens": 1309,
  "midReasoningTokens": 753
}

6680b62a-a3ef-43f6-b85e-e7a5be6ca908

Aug 8, 2026, 09:06:49

togetherai:thinkingmachines/Inkling-Small

Input

Output

375

Reasoning

Answer

375

Cost

$0.000499

Mode

streaming

Cost breakdown

Input $0.000049Output $0.000450

Latency

2.4s

TTFT

174ms

Tok/s

157.4

Tok/s (adj)

169.8

cd67806c-5b59-4a98-a577-b0e692ce3849

Aug 8, 2026, 09:06:56

togetherai:thinkingmachines/Inkling-Small

Input

100

Output

1,309

Reasoning

753

Answer

556

Cost

$0.0016

Mode

streaming

Cost breakdown

Input $0.000050Output $0.0016

Latency

7.0s

TTFT

3.9s

Tok/s

186.9

Tok/s (adj)

179.8

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
}

6fde0d72-596f-47f1-a597-252ef8eb6934

Aug 8, 2026, 09:06:57

togetherai:thinkingmachines/Inkling-Small

Input

130

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000037

Mode

streaming

Cost breakdown

Input $0.000001Output $0.000023Cache read $0.000013

Latency

395ms

Tok/s

48.1

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
}

7fef8f50-d6ba-4987-b125-428a9964c14b

Aug 8, 2026, 09:06:57

togetherai:thinkingmachines/Inkling-Small

Input

132

Output

Reasoning

Answer

Cost

$0.000139

Mode

streaming

Cost breakdown

Input $0.000066Output $0.000073

Latency

455ms

Tok/s

134.1

9fecddcd-c592-403d-a913-bb2ac2ae212d

Aug 8, 2026, 09:06:57

togetherai:thinkingmachines/Inkling-Small

Input

208

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000092

Mode

streaming

Cost breakdown

Input $0.000040Output $0.000040Cache read $0.000013

Latency

314ms

TTFT

245ms

Tok/s

105.1

Tok/s (adj)

262.7

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.

reasoned=true, none token —; follow-up → "The lookup shows **$1,234** for account `acme-42`. Calculation (5% an"

Result

{
  "signatureType": "none",
  "tokenCaptured": false,
  "reasoned": true,
  "answered": true,
  "continuation": "The lookup shows **$1,234** for account `acme-42`.\n\nCalculation (5% annual compound for7 years):\n\n\\[\n\\text{Future Value} =1234 \\times (1.05)"
}

97fd2386-8d88-485e-8a5f-099e0af1046f

Aug 8, 2026, 09:06:58

togetherai:thinkingmachines/Inkling-Small

Input

164

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000085

Mode

streaming

Cost breakdown

Input $0.000018Output $0.000054Cache read $0.000013

Latency

353ms

Tok/s

127.5

b1ce2d33-1a2a-499a-99b2-4af1e215e7c0

Aug 8, 2026, 09:07:01

togetherai:thinkingmachines/Inkling-Small

Input

223

Output

683

Reasoning

499

Answer

184

Cache Read

128

Cost

$0.000880

Mode

streaming

Cost breakdown

Input $0.000048Output $0.000820Cache read $0.000013

Latency

3.5s

TTFT

2.6s

Tok/s

196.2

Tok/s (adj)

207.7

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
}

441673ea-eb7f-40c6-9920-c44feeb7adfc

Aug 8, 2026, 09:07:02

togetherai:thinkingmachines/Inkling-Small

Input

Output

Reasoning

Answer

Cost

$0.000064

Mode

streaming

Cost breakdown

Input $0.000036Output $0.000028

Latency

348ms

TTFT

251ms

Tok/s

66.1

Tok/s (adj)

237.6

summarize-memex

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

Read ~10214 input tokens, wrote 637 words (817 output tokens) in 4907ms.

Result

{
  "wordCount": 637,
  "inputTokens": 10214,
  "outputTokens": 817,
  "latencyMs": 4907
}

51203459-c69f-4f3a-b33b-15eb769c9b8a

Aug 8, 2026, 09:07:06

togetherai:thinkingmachines/Inkling-Small

Input

10,214

Output

817

Reasoning

Answer

817

Cost

$0.0061

Mode

streaming

Cost breakdown

Input $0.0051Output $0.000980

Latency

4.9s

TTFT

535ms

Tok/s

166.5

Tok/s (adj)

186.9

truncation

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

finishReason=length, out=100 tokens.

Result

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

2fe9bd58-e51d-472e-b472-82f7bfe41e64

Aug 8, 2026, 09:07:07

togetherai:thinkingmachines/Inkling-Small

Input

Output

100

Reasoning

Answer

100

Cost

$0.000135

Mode

streaming

Cost breakdown

Input $0.000015Output $0.000120

Latency

772ms

TTFT

256ms

Tok/s

129.5

Tok/s (adj)

193.8

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"
}