LetTokenBurnIt was cited in Anthropic's Fable 5 and Mythos 5 model card.
$ npm i -g lettokenburn-cli$ ltb bench info gdp-pdf
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| GPT-6 Astra | 34.2 | N/A | Independent | medium | N/A | surgehq.ai |
| GPT-5.6 Sol | 30.7 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Opus 5.5 | 30.6 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Fable 5 | 30.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Fable 5.1 | 29.6 | N/A | Independent | medium | N/A | surgehq.ai |
| Muse Spark 1.3 | 27.6 | N/A | Independent | medium | N/A | surgehq.ai |
| GPT-6 Sol | 26.4 | N/A | Independent | medium | N/A | surgehq.ai |
| GPT-5.5 | 26.0 | N/A | Independent | medium | N/A | surgehq.ai |
| GPT-5.6-terra | 24.7 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Opus 4.8 | 24.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Opus 5 | 24.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini 3.7 Flash | 23.8 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini 3.8 Flash | 23.4 | N/A | Independent | medium | N/A | surgehq.ai |
| Qwen3.8 Max | 23.2 | N/A | Independent | medium | N/A | surgehq.ai |
| GPT-6 Luna | 23.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Grok-4.7 | 22.8 | N/A | Independent | medium | N/A | surgehq.ai |
| GPT-5.6 Luna | 22.7 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Opus 4.7 | 21.0 | N/A | Independent | medium | N/A | surgehq.ai |
| DeepSeek V4.1 Flash | 19.8 | N/A | Independent | medium | N/A | surgehq.ai |
| Kimi K3 | 19.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Claude Sonnet 4.6 | 18.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Grok-4.6 | 17.2 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini 3.1 Pro | 17.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Qwen3.8 Flash | 16.6 | N/A | Independent | medium | N/A | surgehq.ai |
| Muse Spark 1.2 | 16.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Muse Spark 1.1 | 15.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini-3.5-flash | 14.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini 3.6 Flash | 14.0 | N/A | Independent | medium | N/A | surgehq.ai |
| GLM-5.3 Flash | 14.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Grok 4.5 | 14.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Kimi K2.6 | 12.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini 3.5 Flash Lite | 10.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Gemini 3 Flash Preview | 10.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Grok-4.3 | 8.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Mistral Large 3 | 2.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Nova 2 | 2.0 | N/A | Independent | medium | N/A | surgehq.ai |
| Nova 2.0 Pro Preview | 2.0 | N/A | Independent | medium | N/A | surgehq.ai |
| NVIDIA Nemotron 3 Nano Omni | 2.0 | N/A | Independent | medium | N/A | surgehq.ai |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
GDP.pdf
GDP.pdf measures professional multimodal reasoning over the documents the economy actually runs on: dense reports, contracts, filings, and records in the messy real-world formats professionals work from. It was cited in Anthropic's Fable 5 and Mythos 5 model card.
What it tests
The benchmark contains 100 real-world prompts and PDFs pulled directly from professional workflows across ten domains: Finance, Healthcare, Legal, STEM/Research, Engineering…