armandmcqueen.dev
← All runs

together:deepseek-v4-pro-0813

8/18/2026, 8:45:52 AM · 15 tests · run 97398d02-e744-402f-8a3f-883168ef87fa

basic-generation

ok

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

  • Produced 94 words (197 output tokens).
Result
{
  "wordCount": 94,
  "outputTokens": 197
}
714343bb-4abc-4ecd-ab85-bf06ea5ec104
Aug 18, 2026, 08:44:53

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
15
Output
197
Reasoning
68
Answer
129
Cost
$0.000800
Mode
streaming
Cost breakdown
Input $0.000020Output $0.000780
Latency
2.6s
TTFT
1.5s
Tok/s
74.8
Tok/s (adj)
118.4

usage-cost

ok

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

  • in=14 out=98 rt=91 cost=$0.000407 (paid).
Result
{
  "inputTokens": 14,
  "outputTokens": 98,
  "reasoningTokens": 91,
  "costUsd": 0.00040655999999999993,
  "paid": true
}
8d6e15af-194c-40be-afb2-7793e79fcb01
Aug 18, 2026, 08:44:56

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
14
Output
98
Reasoning
91
Answer
7
Cost
$0.000407
Mode
streaming
Cost breakdown
Input $0.000018Output $0.000388
Latency
2.6s
TTFT
2.5s
Tok/s
38.0
Tok/s (adj)
63.2

streaming-structure

ok

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

  • mode=streaming, 19 stream events, 1 text block(s), 35 chars.
Result
{
  "mode": "streaming",
  "streamEventCount": 19,
  "textBlocks": 1,
  "textLen": 35
}
5d0421d6-3d56-4f3a-88f0-a613eee5ee51
Aug 18, 2026, 08:44:56

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
17
Output
48
Reasoning
17
Answer
31
Cost
$0.000213
Mode
streaming
Cost breakdown
Input $0.000022Output $0.000190
Latency
743ms
TTFT
605ms
Tok/s
64.6
Tok/s (adj)
224.9

complete-stream-parity

ok

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

  • stream="Paris" · complete="Paris"
Result
{
  "streamText": "Paris",
  "completeText": "Paris"
}
f3864d6e-eecb-4a22-ac11-97dc9a6f0e4c
Aug 18, 2026, 08:44:57

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
18
Output
16
Reasoning
13
Answer
3
Cost
$0.000087
Mode
streaming
Cost breakdown
Input $0.000024Output $0.000063
Latency
628ms
TTFT
524ms
Tok/s
25.5
Tok/s (adj)
28.8
672accc2-9ca0-487b-8eb1-a58b658d76f1
Aug 18, 2026, 08:44:58

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
18
Output
16
Reasoning
13
Answer
3
Cost
$0.000087
Mode
non-streaming
Cost breakdown
Input $0.000024Output $0.000063
Latency
522ms
Tok/s
30.7

reasoning-conformance

ok

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

  • Class (from reasoning-config): always-on.
  • reasoningTokens=337, outputTokens=616.
Result
{
  "class": "always-on",
  "reasoningTokens": 337,
  "outputTokens": 616
}
66f938ba-1fce-475a-990b-4908bfc003a2
Aug 18, 2026, 08:45:11

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
89
Output
616
Reasoning
337
Answer
279
Cost
$0.0026
Mode
streaming
Cost breakdown
Input $0.000117Output $0.0024
Latency
13.5s
TTFT
7.7s
Tok/s
45.8
Tok/s (adj)
48.8

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
}
6a6c8406-487a-485d-8c6b-29385ce832ba
Aug 18, 2026, 08:45:13

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
355
Output
77
Reasoning
16
Answer
61
Cost
$0.000774
Mode
streaming
Cost breakdown
Input $0.000469Output $0.000305
Latency
2.2s
Tok/s
34.9

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: "23 multiplied by 17 is 391."
Result
{
  "toolName": "multiply",
  "args": {
    "a": 23,
    "b": 17
  },
  "product": 391,
  "continuation": "23 multiplied by 17 is 391.",
  "usesResult": true,
  "forced": true
}
7c2f3786-ee13-4fd7-ad91-766fc51a5d51
Aug 18, 2026, 08:45:14

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
351
Output
25
Reasoning
0
Answer
25
Cost
$0.000562
Mode
streaming
Cost breakdown
Input $0.000463Output $0.000099
Latency
1.2s
Tok/s
20.3
6ec13bf4-441b-4456-a1bc-beecff997a29
Aug 18, 2026, 08:45:16

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
427
Output
17
Reasoning
6
Answer
11
Cache Read
351
Cost
$0.000213
Mode
streaming
Cost breakdown
Input $0.000100Output $0.000067Cache read $0.000046
Latency
1.3s
TTFT
1.2s
Tok/s
13.3
Tok/s (adj)
293.3

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.

  • reasoned=true, none token —; follow-up → "The balance for account `acme-42` is **$1,234**. ## Calculation Usin"
Result
{
  "signatureType": "none",
  "tokenCaptured": false,
  "reasoned": true,
  "answered": true,
  "continuation": "The balance for account `acme-42` is **$1,234**.\n\n## Calculation\n\nUsing the compound interest formula:\n\n\\[\nA = P(1 + r)^n\n\\]\n\nWhere:\n- \\(P ="
}
d83e2039-6a07-4191-8812-0751b7d5fdb0
Aug 18, 2026, 08:45:16

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
381
Output
58
Reasoning
8
Answer
50
Cost
$0.000733
Mode
streaming
Cost breakdown
Input $0.000503Output $0.000230
Latency
782ms
Tok/s
74.2
31fe7689-29b3-43dc-b78d-d5cea9bbd7f2
Aug 18, 2026, 08:45:20

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
457
Output
585
Reasoning
377
Answer
208
Cache Read
381
Cost
$0.0025
Mode
streaming
Cost breakdown
Input $0.000100Output $0.0023Cache read $0.000050
Latency
3.8s
TTFT
2.3s
Tok/s
153.5
Tok/s (adj)
135.4

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
}
cf5c6ada-28d4-49af-ad68-64f85f6673e4
Aug 18, 2026, 08:45:29

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
63
Output
110
Reasoning
77
Answer
33
Cost
$0.000519
Mode
streaming
Cost breakdown
Input $0.000083Output $0.000436
Latency
8.5s
TTFT
2.0s
Tok/s
12.9
Tok/s (adj)
5.1

summarize-memex

ok

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

  • Read ~10232 input tokens, wrote 419 words (826 output tokens) in 13586ms.
Result
{
  "wordCount": 419,
  "inputTokens": 10232,
  "outputTokens": 826,
  "latencyMs": 13586
}
02928627-4675-4de6-8fb8-1ad956d92abc
Aug 18, 2026, 08:45:42

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
10,232
Output
826
Reasoning
280
Answer
546
Cost
$0.0168
Mode
streaming
Cost breakdown
Input $0.0135Output $0.0033
Latency
13.6s
TTFT
5.5s
Tok/s
60.8
Tok/s (adj)
67.1

truncation

ok

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

  • finishReason=length, out=100 tokens.
Result
{
  "finishReason": "length",
  "outputTokens": 100
}
381336e2-8638-4c75-b0fc-a766972e8bed
Aug 18, 2026, 08:45:51

togetherai:deepseek-ai/DeepSeek-V4-Pro-0813

Input
21
Output
100
Reasoning
100
Answer
0
Cost
$0.000424
Mode
streaming
Cost breakdown
Input $0.000028Output $0.000396
Latency
8.5s
Tok/s
11.7

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"
}