# Time-to-first-token does not describe a thinking model

> On a Strix Halo box the conventional TTFT figure looked excellent while most requests returned an empty answer with HTTP 200. Client-side measurement, three operating settings, evidence bundles included.

- Published: 2026-08-02
- Evidence level: lab_repeated
- Funding: Self-funded internal research
- Canonical: https://agmind.ai/reports/ttft-thinking-model-strix-halo/

> **Evidence level: `lab_repeated`.** Every number below is re-derived from the
> raw per-request records by the query named in the claim registry — none of them
> is typed into this page. Depth is diagnostic: three measured runs per operating
> setting at concurrency 1, on one unit. This is the pipeline-proof study that
> precedes the flagship, not the flagship itself. Unit-replicated numbers follow
> with the frozen preregistration.

## The short version

A model that reasons before answering breaks the metric everyone quotes.

On the tested configuration, the conventional time-to-first-token looks
excellent — a fifth of a second. But that first token is a *reasoning* token.
The first token of the actual answer arrives roughly a hundred times later, and
under a common token budget most requests never produce an answer at all: the
caller gets HTTP 200 and an empty string.

None of this is visible in a benchmark that reports tokens per second.

## What was measured

| Identity | Value |
|---|---|
| System | Beelink GTR9 Pro (Ryzen AI Max+ 395, Radeon 8060S / gfx1151, 128 GB unified) |
| Runtime | llama.cpp `server-vulkan-b9049`, pinned by image digest |
| Model | Qwen3.6-35B-A3B Q4_K_M from `ggml-org`, downloaded at a pinned revision and hash-verified |
| Workload | `interactive-assistant-v1` @ 2026-08-02 — 16 frozen prompts, EN and RU, three length bands |
| Cell | concurrency 1, context 8192, temperature 0, prompt cache disabled |

Timing is **client-side**: the clock starts when the request leaves and stops on
the tokens as they arrive. Server-reported timings would have hidden exactly the
effect this report is about. Every request carries a unique salt and the prompt
cache is off, so nothing is answered from a previous request's work.

Three operating settings were run, changing one thing at a time:

1. reasoning enabled, 1024-token budget;
2. reasoning enabled, 4096-token budget;
3. reasoning disabled via the chat template, 1024-token budget.

## Results

**The conventional first-token figure is identical in all three settings** —
around **221 ms** ([strix.qwen36.interactive.c1.ttft-any-token](https://agmind.ai/claims/strix.qwen36.interactive.c1.ttft-any-token/)). It cannot
tell apart a configuration that answers immediately from one that never answers.

**With reasoning enabled, the answer starts far later.** At the 4096-token budget
the first *answer* token arrives at a median of
**23453 ms** ([strix.qwen36.interactive.c1.ttfa-thinking](https://agmind.ai/claims/strix.qwen36.interactive.c1.ttfa-thinking/)). The user watches
tokens stream for that entire time — they are the model thinking out loud, not
the reply.

**At the smaller budget, most requests produced no answer at all.** With
reasoning enabled and a 1024-token budget,
**75.0 % of requests** ([strix.qwen36.interactive.c1.answerless-1k](https://agmind.ai/claims/strix.qwen36.interactive.c1.answerless-1k/)) of requests
returned a successful HTTP response with an empty answer: the reasoning pass had
consumed the whole budget before the answer began. An application checking
status codes would have recorded these as successes.

**Disabling reasoning collapses the gap.** With the reasoning block switched off,
the first answer token arrives at
**220 ms** ([strix.qwen36.interactive.c1.ttfa-nothink](https://agmind.ai/claims/strix.qwen36.interactive.c1.ttfa-nothink/)) — the answer starts
with the stream. Median end-to-end time dropped by roughly five times against
the 4096-token reasoning cell, and the answers still passed the language and
repetition gates.

**One gate failed in every setting.** Half the answers exceeded the word budget
declared per prompt. That is a property of the model's verbosity, not of the
reasoning setting, and it is why this configuration would not receive an
unqualified pass on this workload.

## What this means operationally

- **A first-token number without a definition is not usable.** Ask which token:
  the first one streamed, or the first one belonging to the answer. On a
  reasoning model these are different questions with a hundredfold difference in
  the answer.
- **A token budget is a correctness setting, not just a cost setting.** Below the
  budget the reasoning pass needs, the endpoint returns success with nothing in
  it. Anything monitoring only status codes will not notice.
- **Empty-answer responses belong in your error rate.** In this study they are
  counted as failures and stay in the denominator of every latency statistic.

## What this does not establish

Not a capacity claim: concurrency 1 only. Not a quality verdict: the gates cover
format, language and repetition, not correctness. Not a device claim: one unit,
one image digest, one artifact. Repeated but not replicated: three measured runs
per setting on the same unit, which is why the evidence level is `lab_repeated`
and no operating envelope is recommended here. Whether disabling reasoning costs
answer quality is untested — it plausibly does, and that is a separate
qualification.

## Evidence

Each run produced a bundle: run manifest with the full identity, per-request
records including failures, per-request gate outcomes, the runtime log and a
checksum file. The numbers in this report are re-derived from those records by
SQL queries stored beside the claims; the build fails if a rendered value stops
matching its evidence. How that works is described on the
[evidence and data](https://agmind.ai/data/) page.

---

Machine-readable claim registry: https://agmind.ai/claims.json · llms.txt: https://agmind.ai/llms.txt