Claude Mythos Preview95
Not enough families covered yet.Add more benchmark rows for Claude Mythos Preview to see the family profile.
VerdictPal · editorial desk · updated 16 Oct 2026VerdictPal
Models
Model-first evidence atlas.
We score models, not vendors. Below: every model page the public benchmark ledger still has evidence for. The mix is a weighted roll-up of the benchmarks we publish — dated and sourced, not a universal verdict.
234
Models with pages
22
Providers
186
Priced in USD/1M
3
In score ledger only
Intelligence × blended cost
Blended cost per million tokens (input + 3× output) on the X axis, log scale. The Y axis is the benchmark mix from the suites we still publish. Dots are sized by context window and connected by the Pareto frontier.
All 158 models in this chart
Input-only scatter
The earlier cut, using input cost alone. Useful when your workload is input-heavy (RAG, document Q&A) rather than output-heavy.
Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.
AnthropicOpenAIGoogleAlibabaMetaZhipuxAIDeepSeekMiniMaxXiaomiMoonshotOtherMeituanMistralNVIDIACohere
3 models with benchmark rows but no model page
Older releases, special configurations (Codex CLI, ForgeCode), and ad-hoc aggregations. We surface their sourced scores but we don't promote them to the model atlas until they earn a full model page.
Family profile — strongest current evidence rows
The highest intelligence receipt in each family slice. Use this as a sanity check that a single number isn't hiding a weak spot.
Claude Fable 5.177
Family profile
Best published score in each covered benchmark family.
80/100 avg
- Knowledge
- 67
- Reasoning
- 94
- Coding
- 63
- Agentic
- 91
- Long context
- 85
5 tested benchmark families
GPT-6 Astra75
Family profile
Best published score in each covered benchmark family.
81/100 avg
- Reasoning
- 96
- Coding
- 57
- Agentic
- 88
- Long context
- 81
4 tested benchmark families
Which families does the ledger actually cover?
A model with no row in a family is a model that hasn't been publicly tested there. We surface that instead of pretending it's weak.
Knowledge13models covered
Reasoning194models covered
Math15models covered
Coding98models covered
Agentic168models covered
Multimodal8models covered
Long context184models covered
Tool use0models covered
Performance0models covered
Safety0models covered
Human preference0models covered
Where these models show up in the atlas
Every tool on this list is wired to at least one model in the atlas above. The same wrapper that hides the harness can also hide the price.
Questions this atlas answers
Every answer below is assembled from the dated fields on this page. Nothing is written separately for search.
What is the VerdictPal model atlas?
We score models, not vendors. Below: every model page the public benchmark ledger still has evidence for. The mix is a weighted roll-up of the benchmarks we publish — dated and sourced, not a universal verdict.
How many models have pages?
234 models currently have a page in the atlas. Draft stubs are omitted.
Where do the intelligence rows come from?
From the benchmarks we still publish. Every row on this page carries a date and a URL. VerdictPal does not invent model scores.
Are model prices dated?
186 of 234 models publish a per-token USD price. Each model page records the date that price was checked.
The Pack · Editorial newsletter
Know when the rules move.
One short email when a university changes its AI rule for theses, or a research tool changes its pricing or privacy terms. No tracking, no third-party analytics. Unsubscribe in one click.