# TubeLab Scriptwriting > Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/tubelab%3Ascriptwriting. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/tubelab%3Ascriptwriting. Writing long-form video scripts for twelve test prompts, assessed by blind model judges on hook, retention, storytelling, voice, pacing and clarity; reported cost is relative to a fixed baseline. Canonical URL: https://benchize.com/benchmarks/tubelab%3Ascriptwriting Last updated: Wed, 21 Oct 2026 18:52:20 GMT Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations. Writing long-form video scripts for twelve test prompts, assessed by blind model judges on hook, retention, storytelling, voice, pacing and clarity; reported cost is relative to a fixed baseline. Writing & creativity · Models tested: 16 · Score publishers: [tubelab](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) ## Model results | # | Model | Score | Cost | | | --- | --- | --- | --- | --- | | 1 | [Claude Opus 5](https://benchize.com/models/claude-opus-5.md) | 87.9% | 1.82ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 87.9 \| percent \| source_reported \| Higher is better \| \| `craft` \| 87.9 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 1.82 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 8.3 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 8.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 8.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 8.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 9.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":89.7,"low":85.8} headToHead {"losses":2,"ties":3,"winPercent":65,"wins":7} rank 1 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Claude Opus 5 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/0` | | | | | | 2 | [Claude Fable 5.1](https://benchize.com/models/claude-fable-5-1.md) | 85.7% | 2.69ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 85.7 \| percent \| source_reported \| Higher is better \| \| `craft` \| 85.7 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 2.69 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 6.7 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 8.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 8.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 8.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 9.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":89.5,"low":81.4} headToHead {"losses":2,"ties":3,"winPercent":65,"wins":7} rank 2 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Claude Fable 5.1 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/1` | | | | | | 3 | [Claude Opus 4.8](https://benchize.com/models/claude-opus-4-8.md) | 84.4% | 1.4ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 84.4 \| percent \| source_reported \| Higher is better \| \| `craft` \| 84.4 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 1.4 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 6.4 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 8.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 9.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":87.8,"low":80.5} headToHead {"losses":5,"ties":2,"winPercent":52,"wins":5} rank 3 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Claude Opus 4.8 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/2` | | | | | | 4 | [Claude Fable 5](https://benchize.com/models/claude-fable-5.md) | 82.2% | 2.99ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 82.2 \| percent \| source_reported \| Higher is better \| \| `craft` \| 87.4 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 2.99 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 7.5 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 8.6 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 8.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 8.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 9.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":88.4,"low":74.4} headToHead {"losses":2,"ties":2,"winPercent":65,"wins":6} rank 7 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Claude Fable 5 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/6` | | | | | | 5 | [GPT-5.5](https://benchize.com/models/gpt-5-5.md) | 80.3% | 1.13ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 80.3 \| percent \| source_reported \| Higher is better \| \| `craft` \| 80.3 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 1.13 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 4.7 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 9.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":85.4,"low":74} headToHead {"losses":4,"ties":1,"winPercent":56,"wins":7} rank 9 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label GPT-5.5 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/8` | | | | | | 6 | [Claude Sonnet 5](https://benchize.com/models/claude-sonnet-5.md) | 79.3% | 0.71ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 79.3 \| percent \| source_reported \| Higher is better \| \| `craft` \| 79.3 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.71 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 7.5 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":84.8,"low":73.3} headToHead {"losses":4,"ties":3,"winPercent":45,"wins":3} rank 10 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Claude Sonnet 5 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/9` | | | | | | 7 | [Muse Spark 1.1](https://benchize.com/models/muse-spark-1-1.md) | 79% | 1.09ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 79 \| percent \| source_reported \| Higher is better \| \| `craft` \| 79 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 1.09 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 4.5 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 8.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":84.2,"low":72.6} headToHead {"losses":7,"ties":2,"winPercent":28,"wins":1} rank 11 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Muse Spark 1.1 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/10` | | | | | | 8 | [Gemini 3.8 Flash](https://benchize.com/models/gemini-3-8-flash.md) | 78.5% | 0.17ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 78.5 \| percent \| source_reported \| Higher is better \| \| `craft` \| 78.5 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.17 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 1.8 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.4 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.6 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":80.6,"low":76.2} headToHead {"losses":8,"ties":2,"winPercent":27,"wins":2} rank 12 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Gemini 3.8 Flash Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/11` | | | | | | 9 | [DeepSeek V4 Pro 0813](https://benchize.com/models/deepseek-v4-pro-0813.md) | 76.5% | 0.39ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 76.5 \| percent \| source_reported \| Higher is better \| \| `craft` \| 76.5 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.39 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 5.3 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 7.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":81.4,"low":70.7} headToHead {"losses":6,"ties":1,"winPercent":41,"wins":4} rank 14 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label DeepSeek V4 Pro (0813) Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/13` | | | | | | 10 | [DeepSeek V4 Pro](https://benchize.com/models/deepseek-v4-pro.md) | 76.4% | 0.05ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 76.4 \| percent \| source_reported \| Higher is better \| \| `craft` \| 76.4 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.05 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 10.6 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 7.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.6 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":81.4,"low":71.5} headToHead {"losses":6,"ties":1,"winPercent":42,"wins":4} rank 15 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label DeepSeek V4 Pro Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/14` | | | | | | 11 | [DeepSeek V4 Flash](https://benchize.com/models/deepseek-v4-flash.md) | 75.9% | 0.05ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 75.9 \| percent \| source_reported \| Higher is better \| \| `craft` \| 77.7 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.05 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 8.9 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 7.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.6 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 7.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":80.7,"low":70.1} headToHead {"losses":5,"ties":3,"winPercent":42,"wins":2} rank 16 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label DeepSeek V4 Flash Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/15` | | | | | | 12 | [Claude Sonnet 4.6](https://benchize.com/models/claude-sonnet-4-6.md) | 75.4% | 0.7ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 75.4 \| percent \| source_reported \| Higher is better \| \| `craft` \| 75.4 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.7 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 7.3 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 6.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 6.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 9 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":80.6,"low":70.1} headToHead {"losses":7,"ties":2,"winPercent":27,"wins":2} rank 17 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Claude Sonnet 4.6 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/16` | | | | | | 13 | [GPT-5.6 Sol](https://benchize.com/models/gpt-5-6-sol.md) | 75.3% | 1.02ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 75.3 \| percent \| source_reported \| Higher is better \| \| `craft` \| 75 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 1.02 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 4.1 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 7.1 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 6.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 6.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":80.3,"low":69.8} headToHead {"losses":4,"ties":3,"winPercent":52,"wins":4} rank 18 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label GPT-5.6 Sol Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/17` | | | | | | 14 | [Gemini 3.6 Flash](https://benchize.com/models/gemini-3-6-flash.md) | 75% | 0.43ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 75 \| percent \| source_reported \| Higher is better \| \| `craft` \| 75 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.43 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 3.9 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 6.9 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":78.8,"low":71} headToHead {"losses":9,"ties":1,"winPercent":21,"wins":2} rank 19 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Gemini 3.6 Flash Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/18` | | | | | | 15 | [Muse Spark 1.3](https://benchize.com/models/muse-spark-1-3.md) | 72.8% | 0.29ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 72.8 \| percent \| source_reported \| Higher is better \| \| `craft` \| 72.8 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 0.29 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 4.3 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 7.1 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 6.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 7.1 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 7.2 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 8.3 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":78.8,"low":66.6} headToHead {"losses":8,"ties":0,"winPercent":35,"wins":4} rank 20 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label Muse Spark 1.3 Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/19` | | | | | | 16 | [GPT-6 Astra](https://benchize.com/models/gpt-6-astra.md) | 56.7% | 1.78ratio | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `quality` \| 56.7 \| percent \| source_reported \| Higher is better \| \| `craft` \| 56.7 \| percent \| source_reported \| Higher is better \| \| `relative_cost_per_script` \| 1.78 \| ratio \| source_reported \| Lower is better \| \| `script_time` \| 2.8 \| minutes \| source_reported \| Lower is better \| \| `hook` \| 5.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `retention` \| 5.7 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `storytelling` \| 4.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `voiceMatch` \| 5 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `pacing` \| 5.1 \| score_out_of_10 \| craft_rubric \| Higher is better \| \| `clarity` \| 7.8 \| score_out_of_10 \| craft_rubric \| Higher is better \| ### Configuration judgingProtocol blind_model_judges Reasoning effort unknown taskCount 12 ### Reported statistics confidence {"high":64.7,"low":49} headToHead {"losses":12,"ties":0,"winPercent":0,"wins":0} rank 22 ### Source [Original source](https://tubelab.net/_next/static/chunks/pages/benchmark/scriptwriting-7ad7196ff3093ad4.js) Source tubelab Original model label GPT-6 Astra Observed 2026-10-21T16:34:03.856Z Evidence SHA-256 `72f9d867f388256af8fc6dca12e8c7b6f1ddcbb97f0712da339b73102686761f` Source locator `/rows/21` | | | | | --- Cite as: TubeLab Scriptwriting — Benchize, https://benchize.com/benchmarks/tubelab%3Ascriptwriting (retrieved 2026-10-21)