[whatmodel.sh](../../README.md) / Claude Mythos Preview (limited preview) # Claude Mythos Preview (limited preview) Anthropic · released 2026-04 · proprietary **score 79.3** ±1.1 · #15 of 732 · from 12 benchmark results across 1 source · † no independently run results verified Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at ● marks Claude Mythos Preview (limited preview) among every model on that task or test, worst to best. 
task                vs the others         rank  counts  tests
agent tasks         ▁▁▃▆█▄▅▅▄▄●▁    #13 of 313     68%      7
coding              ▁▁▁▃▅▇█▇▅▄●▁    #12 of 469     45%      3
math                ▁▁▃▅█▆▄█▄▂●▁   #11 of 431*    2.2%      1
vision & documents  ▁▁▂▂▁▃▄▂▆▇█●    #1 of 153*    5.7%      1
web search          ▁▁▁▅▄█●▄·▁·▁    #23 of 84*     12%      2
agent tasks · #13 of 313 · 68% of its score Beats [GPT-5.5 Pro](gpt-5-5-pro.md), [GPT-5.6 Sol](gpt-5-6-sol.md) here, though they rank higher overall. 
test                                result  vs the others in its run    difficulty            counts  why
SWE-bench Pro (split unspecified)…   77.8%  ▄█·▄·●██···▄ beats 4/9      very hard · avg 7.4%     24%  sharp
SWE-bench Multilingual (maker)       87.3%  ▃··▃·▅▃●█··▃ beats 5/9      hard · avg 31%           11%  typical
SWE-bench Verified (maker)           93.9%  ▁▁▁▁▁▁▂▄█▆▁● beats 136/138  hard · avg 28%          9.4%  sharp · maxed out
Humanity's Last Exam with tools (…   64.7%  ▁······▁·▄·● beats 8/10     very hard · avg 14%     7.4%  at its level
OSWorld-Verified (maker)             85.4%  █▃·▂▅▂▃▂▅▃▆● beats 27/27    hard · avg 28%          6.7%  typical
Terminal-Bench 2.0 (maker)           82.0%  ▂▂▅▅████▆▂▃● beats 35/36    very hard · avg 14%     4.8%  small run
BrowseComp with context managemen…   86.9%  ▄·▄▂·█▂·▆●·▂ beats 13/15    medium · avg 47%        4.3%  typical
coding · #12 of 469 · 45% of its score Beats [GPT-6.1 Terra](gpt-6-1-terra.md), [GPT-5.6 Sol](gpt-5-6-sol.md), [Claude Fable 5](claude-fable-5.md) here, though they rank higher overall. Trails [Qwen3.8 Max (0902)](qwen3-8-max-0902.md) here, though they rank lower overall. 
test                                result  vs the others in its run    difficulty            counts  why
SWE-bench Pro (split unspecified)…   77.8%  ▄█·▄·●██···▄ beats 4/9      very hard · avg 7.4%     24%  sharp
SWE-bench Multilingual (maker)       87.3%  ▃··▃·▅▃●█··▃ beats 5/9      hard · avg 31%           11%  typical
SWE-bench Verified (maker)           93.9%  ▁▁▁▁▁▁▂▄█▆▁● beats 136/138  hard · avg 28%          9.4%  sharp · maxed out
math · #11 of 431 · 2.2% of its score Beats [Claude Opus 5](claude-opus-5.md) here, though they rank higher overall. 
test                result  vs the others in its run  difficulty      counts  why
USAMO 2026 (maker)   97.6%  █·██······█● beats 4/4    hard · avg 19%    2.2%  sharp · noisy
vision & documents · #1 of 153 · 5.7% of its score Beats [GPT-6 Astra](gpt-6-astra.md), [Claude Opus 5.5](claude-opus-5-5.md), [GPT-6.1 Sol](gpt-6-1-sol.md) here, though they rank higher overall. 
test                                result  vs the others in its run  difficulty      counts  why
CharXiv (reasoning questions) (ma…   86.2%  ███··█·██··● beats 6/6    hard · avg 28%    5.7%  typical
web search · #23 of 84 · 12% of its score Trails [GPT-4o mini (2024-07-18)](gpt-4o-mini-2024-07-18.md), [Mistral Small 3.1 24B (2503)](mistral-small-3-1.md), [GPT-4o (2024-08-06)](gpt-4o-2024-08-06.md) here, though they rank lower overall. 
test                                result  vs the others in its run  difficulty           counts  why
Humanity's Last Exam with tools (…   64.7%  ▁······▁·▄·● beats 8/10   very hard · avg 14%    7.4%  at its level
BrowseComp with context managemen…   86.9%  ▄·▄▂·█▂·▆●·▂ beats 13/15  medium · avg 47%       4.3%  typical
other tests (3) 
test                          result  vs the others in its run    difficulty            counts  why
Humanity's Last Exam (maker)   56.8%  ▂·····▂·▄▂●▄ beats 6/10     very hard · avg 5.9%     14%  at its level
GPQA Diamond (maker)           94.5%  ▁▁▁▁▂▄▅▅███● beats 316/322  hard · avg 39%          7.0%  sharp · maxed out · small run
MMMLU (maker)                  92.7%  ▂▁▁▂▃▂▄▄▅██● beats 46/46    medium · avg 61%        3.0%  small run
## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its *counts* column above; here it is added up by who ran the tests. [Method](../method.md) 
source                                    results  share of the score
Anthropic system cards and model announ…       12   99% ▇▇▇▇▇▇▇▇▇▇▇▇
prior: pull toward the middle                        1% ▇
weighted average                                   = 79.3
## How to run it No verified API or local download yet. [interactive version](https://whatmodel.sh/model/claude-mythos-preview) · [JSON](https://whatmodel.sh/api?format=json&all=1)