grok-4.6

8/19/2026, 7:48:57 AM · 15 tests · run 98a615d9-8179-4ec8-b865-7fe2f693fd2f

basic-generation

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

Produced 94 words (664 output tokens).

Result

{
  "wordCount": 94,
  "outputTokens": 664
}

e0067f97-a84c-4564-a7c2-c58519c4ab23

Aug 19, 2026, 07:47:29

xai:grok-4.6

Input

217

Output

664

Reasoning

543

Answer

121

Cache Read

128

Cost

$0.0042

Mode

streaming

Cost breakdown

Input $0.000178Output $0.0040Cache read $0.000064

Latency

12.0s

TTFT

10.5s

Tok/s

55.5

Tok/s (adj)

83.6

usage-cost

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

in=215 out=255 rt=250 cost=$0.001768 (paid).

Result

{
  "inputTokens": 215,
  "outputTokens": 255,
  "reasoningTokens": 250,
  "costUsd": 0.0017680000000000003,
  "paid": true
}

2b70cd7a-e70d-447a-a2a4-d1a8b4a6d77a

Aug 19, 2026, 07:47:35

xai:grok-4.6

Input

215

Output

255

Reasoning

250

Answer

Cache Read

128

Cost

$0.0018

Mode

streaming

Cost breakdown

Input $0.000174Output $0.0015Cache read $0.000064

Latency

5.7s

TTFT

5.6s

Tok/s

44.6

Tok/s (adj)

60.7

streaming-structure

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

mode=streaming, 56 stream events, 1 text block(s), 35 chars.

Result

{
  "mode": "streaming",
  "streamEventCount": 56,
  "textBlocks": 1,
  "textLen": 35
}

cf6ed584-20f4-465a-862c-e76536da00f4

Aug 19, 2026, 07:47:37

xai:grok-4.6

Input

219

Output

130

Reasoning

101

Answer

Cache Read

128

Cost

$0.0010

Mode

streaming

Cost breakdown

Input $0.000182Output $0.000780Cache read $0.000064

Latency

2.4s

TTFT

2.1s

Tok/s

53.1

Tok/s (adj)

79.8

complete-stream-parity

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

stream="Paris" · complete="Paris"

Result

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

2a9a8199-aeb9-49bf-8861-d5499c17dfe6

Aug 19, 2026, 07:47:39

xai:grok-4.6

Input

220

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000680

Mode

streaming

Cost breakdown

Input $0.000184Output $0.000432Cache read $0.000064

Latency

2.0s

TTFT

2.0s

Tok/s

36.4

Tok/s (adj)

47.7

aad13672-c3c1-4230-85cf-70b12e53ed0d

Aug 19, 2026, 07:47:40

xai:grok-4.6

Input

220

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000572

Mode

non-streaming

Cost breakdown

Input $0.000184Output $0.000324Cache read $0.000064

Latency

1.3s

Tok/s

42.2

reasoning-conformance

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

Class (from reasoning-config): always-on.
reasoningTokens=421, outputTokens=638.

Result

{
  "class": "always-on",
  "reasoningTokens": 421,
  "outputTokens": 638
}

39bcb78f-cdb4-41a8-8a08-76435ade2de3

Aug 19, 2026, 07:47:51

xai:grok-4.6

Input

291

Output

638

Reasoning

421

Answer

217

Cache Read

128

Cost

$0.0042

Mode

streaming

Cost breakdown

Input $0.000326Output $0.0038Cache read $0.000064

Latency

10.6s

TTFT

6.9s

Tok/s

60.4

Tok/s (adj)

59.8

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
}

b570c1e4-afc7-4a00-856c-ac8cca615a13

Aug 19, 2026, 07:47:53

xai:grok-4.6

Input

358

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000944

Mode

streaming

Cost breakdown

Input $0.000460Output $0.000420Cache read $0.000064

Latency

1.8s

Tok/s

40.0

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 of 23 and 17 is 391."

Result

{
  "toolName": "multiply",
  "args": {
    "a": 23,
    "b": 17
  },
  "product": 391,
  "continuation": "The product of 23 and 17 is 391.",
  "usesResult": true,
  "forced": true
}

a4b6fbfc-ebe7-430f-93ce-70ebe4248122

Aug 19, 2026, 07:47:58

xai:grok-4.6

Input

365

Output

239

Reasoning

225

Answer

Cache Read

128

Cost

$0.0020

Mode

streaming

Cost breakdown

Input $0.000474Output $0.0014Cache read $0.000064

Latency

5.9s

Tok/s

40.6

f722cee2-66fb-4717-9777-b136b49d0de4

Aug 19, 2026, 07:48:00

xai:grok-4.6

Input

377

Output

Reasoning

Answer

Cache Read

128

Cost

$0.000634

Mode

streaming

Cost breakdown

Input $0.000498Output $0.000072Cache read $0.000064

Latency

1.1s

TTFT

947ms

Tok/s

10.6

Tok/s (adj)

65.1

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 returned a current balance of $P=1234$ USD. After 7 year"

Result

{
  "signatureType": "none",
  "tokenCaptured": false,
  "reasoned": true,
  "answered": true,
  "continuation": "The lookup returned a current balance of \\(P=1234\\) USD.\n\nAfter 7 years of 5% annual compounding the future value is\n\\[\nA=P(1.05)^7=1234\\tim"
}

06ca38e9-5d45-4320-8942-48ad104fb698

Aug 19, 2026, 07:48:01

xai:grok-4.6

Input

385

Output

Reasoning

Answer

Cache Read

128

Cost

$0.0011

Mode

streaming

Cost breakdown

Input $0.000514Output $0.000540Cache read $0.000064

Latency

1.8s

Tok/s

49.0

808b4fae-54f1-4b1a-a06a-b0f6ec01e902

Aug 19, 2026, 07:48:12

xai:grok-4.6

Input

412

Output

640

Reasoning

532

Answer

108

Cache Read

128

Cost

$0.0045

Mode

streaming

Cost breakdown

Input $0.000568Output $0.0038Cache read $0.000064

Latency

10.4s

TTFT

8.5s

Tok/s

61.5

Tok/s (adj)

57.9

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"
  },
  "validJson": true,
  "missingKeys": [],
  "bareJson": true
}

0f4eca2d-1a82-4581-914c-1d3cee9f05d0

Aug 19, 2026, 07:48:14

xai:grok-4.6

Input

265

Output

114

Reasoning

Answer

Cache Read

128

Cost

$0.0010

Mode

streaming

Cost breakdown

Input $0.000274Output $0.000684Cache read $0.000064

Latency

2.4s

TTFT

2.2s

Tok/s

46.8

Tok/s (adj)

78.6

summarize-memex

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

Read ~10207 input tokens, wrote 451 words (887 output tokens) in 23157ms.

Result

{
  "wordCount": 451,
  "inputTokens": 10207,
  "outputTokens": 887,
  "latencyMs": 23157
}

8a705835-0e62-4b2f-bd23-3c239aeef15b

Aug 19, 2026, 07:48:37

xai:grok-4.6

Input

10,207

Output

887

Reasoning

311

Answer

576

Cache Read

128

Cost

$0.0255

Mode

streaming

Cost breakdown

Input $0.0202Output $0.0053Cache read $0.000064

Latency

23.2s

TTFT

8.0s

Tok/s

38.3

Tok/s (adj)

37.9

truncation

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

finishReason=length, out=615 tokens.

Result

{
  "finishReason": "length",
  "outputTokens": 615
}

64f29f4a-3217-4625-a834-7674d19c2dca

Aug 19, 2026, 07:48:53

xai:grok-4.6

Input

224

Output

615

Reasoning

515

Answer

100

Cache Read

128

Cost

$0.0039

Mode

streaming

Cost breakdown

Input $0.000192Output $0.0037Cache read $0.000064

Latency

15.5s

TTFT

13.1s

Tok/s

39.8

Tok/s (adj)

43.2

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"
}