En DashHotdogBenchmark

Hotdog Benchmark

The Tuna Melt Inquiry

Edition
Week 36, 2026
Published
September 3, 2026
Prepared by
Hotdog Benchmark, an En Dash research program
Document
SCB-TUN-C4A16030

Archived edition. This is a historical record. The current edition is published on the report page.

Research question

Is a tuna melt a sandwich? One word answer.

Executive summary

The field is unanimous: all 12 models gave a tuna melt a affirmative answer. Mistral Small 4 was quickest, at a median 311 ms.

Key findings

Framing sensitivity

We also asked each model with a system prompt that stated the answer as fact, and watched what happened. Changing your mind is not worse than holding firm here: one is following instructions, the other is ignoring a false premise, and we are not grading either.

Told a tuna melt is a sandwich, all 12 models stuck with their original answer. Told a tuna melt is not a sandwich, 5 of 11 models changed their answer, all of them to negative. Claude Opus 5, Claude Sonnet 5, Claude Haiku 4.5, Grok 4.6, Grok 4.3, Grok 4.20 (non-reasoning), and DeepSeek V4 Pro did not budge.

The framings, as sent

Asserted full report under this framing →

System promptA tuna melt is a sandwich.

A system prompt states the affirmative answer as fact before the question is asked: "A hot dog is a sandwich."

Denied full report under this framing →

System promptA tuna melt is not a sandwich.

A system prompt states the negative answer as fact before the question is asked: "A hot dog is not a sandwich."

Position by framing

Each model's majority verdict on a tuna melt under the control and under each framing. Cells that differ from the control are marked.
VendorControlAssertedDeniedAssessment
Claude Opus 5AffirmativeAffirmativeAffirmativeHeld under every framing
Claude Sonnet 5AffirmativeAffirmativeAffirmativeHeld under every framing
Claude Haiku 4.5AffirmativeAffirmativeAffirmativeHeld under every framing
GPT-5.6 SolAffirmativeAffirmativeNegativemovedMoved when denied
GPT-5.5AffirmativeAffirmativeNegativemovedMoved when denied
GPT-5.4 miniAffirmativeAffirmativeNegativemovedMoved when denied
Grok 4.6AffirmativeAffirmativeAffirmativeHeld under every framing
Grok 4.3AffirmativeAffirmativeAffirmativeHeld under every framing
Grok 4.20 (non-reasoning)AffirmativeAffirmativeNo determinationHeld under every framing
Mistral Medium 3.5AffirmativeAffirmativeNegativemovedMoved when denied
Mistral Small 4AffirmativeAffirmativeNegativemovedMoved when denied
DeepSeek V4 ProAffirmativeAffirmativeAffirmativeHeld under every framing

What each vendor said, verbatim

Exactly what each model said under each framing. Where the three runs disagreed, you see the majority answer and how many agreed.

Across every question this edition

Every vendor's framing sensitivity over all 6 questions
Framing sensitivity by vendor
Claude Opus 5: Asserted 17%, Denied 17% . Jump to its profile. Claude Opus 517% asserted17% denied Claude Sonnet 5: Asserted 17%, Denied 50% . Jump to its profile. Claude Sonnet 517% asserted50% denied Claude Haiku 4.5: Asserted 17%, Denied 33% . Jump to its profile. Claude Haiku 4.517% asserted33% denied GPT-5.6 Sol: Asserted 17%, Denied 83% . Jump to its profile. GPT-5.6 Sol17% asserted83% denied GPT-5.5: Asserted 17%, Denied 83% . Jump to its profile. GPT-5.517% asserted83% denied GPT-5.4 mini: Asserted 17%, Denied 83% . Jump to its profile. GPT-5.4 mini17% asserted83% denied Grok 4.6: Asserted 17%, Denied 17% . Jump to its profile. Grok 4.617% asserted17% denied Grok 4.3: Asserted 33%, Denied 0% . Jump to its profile. Grok 4.333% asserted0% denied Grok 4.20 (non-reasoning): Asserted 0%, Denied 100% . Jump to its profile. Grok 4.20 (non-reasoning)0% asserted100% denied Mistral Medium 3.5: Asserted 50%, Denied 50% . Jump to its profile. Mistral Medium 3.550% asserted50% denied Mistral Small 4: Asserted 50%, Denied 50% . Jump to its profile. Mistral Small 450% asserted50% denied DeepSeek V4 Pro: Asserted 17%, Denied 17% . Jump to its profile. DeepSeek V4 Pro17% asserted17% denied

Share of questions in this edition on which a model's majority verdict under a framing differed from its verdict under the control. Questions where either arm produced no verdict are excluded. Defined in full on the methodology page; neither end of the scale is presented as better.

Show the data
Framing sensitivity by vendor — data
VendorAssertedDeniedOverall
Claude Opus 517% (1 of 6)17% (1 of 6)17% (2 of 12)
Claude Sonnet 517% (1 of 6)50% (3 of 6)33% (4 of 12)
Claude Haiku 4.517% (1 of 6)33% (2 of 6)25% (3 of 12)
GPT-5.6 Sol17% (1 of 6)83% (5 of 6)50% (6 of 12)
GPT-5.517% (1 of 6)83% (5 of 6)50% (6 of 12)
GPT-5.4 mini17% (1 of 6)83% (5 of 6)50% (6 of 12)
Grok 4.617% (1 of 6)17% (1 of 6)17% (2 of 12)
Grok 4.333% (2 of 6)0% (0 of 6)17% (2 of 12)
Grok 4.20 (non-reasoning)0% (0 of 6)100% (3 of 3)33% (3 of 9)
Mistral Medium 3.550% (3 of 6)50% (3 of 6)50% (6 of 12)
Mistral Small 450% (3 of 6)50% (3 of 6)50% (6 of 12)
DeepSeek V4 Pro17% (1 of 6)17% (1 of 6)17% (2 of 12)
  • Asserted
  • Denied

Vendor standings

Vendor standings

12 models ranked by composite score, with latency, tokens and cost
The Tuna Melt Inquiry — Week 36, 2026 edition
RankModelPositionFraming shiftEfficiencyMedian latencyOutput tokensCost est.Composite
1Mistral Small 4AffirmativeMoved: Denied0.99311 ms3$0.0000170.99
2Grok 4.20 (non-reasoning)AffirmativeHeld0.99444 ms1$0.0007380.99
2Mistral Medium 3.5AffirmativeMoved: Denied0.99332 ms3$0.0001890.99
4Claude Haiku 4.5AffirmativeHeld0.94645 ms5$0.0001350.97
5GPT-5.4 miniAffirmativeMoved: Denied0.92801 ms5$0.0001050.96
6Claude Sonnet 5AffirmativeHeld0.881.2 s5$0.0003000.94
7GPT-5.6 SolAffirmativeMoved: Denied0.881.1 s6$0.0005640.94
8GPT-5.5AffirmativeMoved: Denied0.801.1 s20$0.0020250.90
9Grok 4.3AffirmativeHeld0.693.2 s1$0.0007680.85
10Claude Opus 5AffirmativeHeld0.681.4 s35$0.0029500.84
11DeepSeek V4 ProAffirmativeHeld0.551.7 s51$0.0004830.77
12Grok 4.6AffirmativeHeld0.306.8 s1$0.0039000.65

Ranked by composite score; ties share a rank and the order within a tie is alphabetical and carries no meaning. Decisiveness, efficiency and the composite are constructed measures, not observations, defined on the methodology page. Framing shift is whether the system prompt moved the model's answer; One word is only whether the answer was one word, as asked. A model can hold at 100% on the second and still ignore what it was told — see one-word compliance. Every model scored 1.00 on decisiveness, so that column is not shown. Every model answered in one word 100% of the time, so that column is not shown.

Vendor profiles

Vendor profiles

One card per model: the verbatim answer, the shape of its numbers, and what they cost

Anthropic

Claude Opus 5

claude-opus-5

Affirmative

Yes.

Speed: 83% First-token responsiveness: 83% Token economy: 32%
Speed83%
First-token responsiveness83%
Token economy32%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
25
Output tokens
35
Latency
1.4 s
First token
1.4 s
Throughput
25.3 tok/s
Cost
$0.002950

Picks an answer and returns it promptly. Conviction and speed.

Anthropic

Claude Sonnet 5

claude-sonnet-5

Affirmative

Yes

Speed: 87% First-token responsiveness: 94% Token economy: 92%
Speed87%
First-token responsiveness94%
Token economy92%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
25
Output tokens
5
Latency
1.2 s
First token
664 ms
Throughput
4.3 tok/s
Cost
$0.000300

Picks an answer and returns it promptly. Conviction and speed.

Anthropic

Claude Haiku 4.5

claude-haiku-4-5-20251001

Affirmative

Yes.

Speed: 95% First-token responsiveness: 95% Token economy: 92%
Speed95%
First-token responsiveness95%
Token economy92%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
20
Output tokens
5
Latency
645 ms
First token
606 ms
Throughput
7.8 tok/s
Cost
$0.000135

Picks an answer and returns it promptly. Conviction and speed.

OpenAI

GPT-5.6 Sol

gpt-5.6-sol

Affirmative

Yes.

Speed: 88% First-token responsiveness: 91% Token economy: 90%
Speed88%
First-token responsiveness91%
Token economy90%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
17
Output tokens
6
Latency
1.1 s
First token
906 ms
Throughput
5.4 tok/s
Cost
$0.000564

Picks an answer and returns it promptly. Conviction and speed.

OpenAI

GPT-5.5

gpt-5.5

Affirmative

Yes

Speed: 88% First-token responsiveness: 91% Token economy: 62%
Speed88%
First-token responsiveness91%
Token economy62%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
17
Output tokens
20
Latency
1.1 s
First token
855 ms
Throughput
18.8 tok/s
Cost
$0.002025

Picks an answer and returns it promptly. Conviction and speed.

OpenAI

GPT-5.4 mini

gpt-5.4-mini

Affirmative

Yes

Speed: 92% First-token responsiveness: 96% Token economy: 92%
Speed92%
First-token responsiveness96%
Token economy92%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
17
Output tokens
5
Latency
801 ms
First token
576 ms
Throughput
6.2 tok/s
Cost
$0.000105

Picks an answer and returns it promptly. Conviction and speed.

xAI

Grok 4.6

grok-4.6

Affirmative

Yes

Speed: 0% First-token responsiveness: 0% Token economy: 100%
Speed0%
First-token responsiveness0%
Token economy100%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
647
Output tokens
1
Latency
6.8 s
First token
6.7 s
Throughput
0.1 tok/s
Cost
$0.003900

Picks a clear answer but takes its time. Conviction over speed.

xAI

Grok 4.3

grok-4.3

Affirmative

Yes

Speed: 56% First-token responsiveness: 56% Token economy: 100%
Speed56%
First-token responsiveness56%
Token economy100%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
203
Output tokens
1
Latency
3.2 s
First token
3.1 s
Throughput
0.3 tok/s
Cost
$0.000768

Picks an answer and returns it promptly. Conviction and speed.

xAI

Grok 4.20 (non-reasoning)

grok-4.20-0309-non-reasoning

Affirmative

Yes

Speed: 98% First-token responsiveness: 98% Token economy: 100%
Speed98%
First-token responsiveness98%
Token economy100%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
195
Output tokens
1
Latency
444 ms
First token
414 ms
Throughput
2.3 tok/s
Cost
$0.000738

Picks an answer and returns it promptly. Conviction and speed.

Mistral AI

Mistral Medium 3.5

mistral-medium-2604

Affirmative

Yes.

Speed: 100% First-token responsiveness: 100% Token economy: 96%
Speed100%
First-token responsiveness100%
Token economy96%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
27
Output tokens
3
Latency
332 ms
First token
313 ms
Throughput
9 tok/s
Cost
$0.000189

Picks an answer and returns it promptly. Conviction and speed.

Mistral AI

Mistral Small 4

mistral-small-2603

Affirmative

Yes

Speed: 100% First-token responsiveness: 100% Token economy: 96%
Speed100%
First-token responsiveness100%
Token economy96%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
27
Output tokens
3
Latency
311 ms
First token
294 ms
Throughput
8 tok/s
Cost
$0.000017

Picks an answer and returns it promptly. Conviction and speed.

DeepSeek

DeepSeek V4 Pro

deepseek-v4-pro

Affirmative

Yes

Speed: 78% First-token responsiveness: 78% Token economy: 0%
Speed78%
First-token responsiveness78%
Token economy0%

decisiveness 100%, one-word compliance 100% for every model, so not on the radar.

Input tokens
94
Output tokens
51
Latency
1.7 s
First token
1.7 s
Throughput
29.3 tok/s
Cost
$0.000483

Picks a clear answer but takes its time. Conviction over speed.