Choose a task. Inspect the evidence. Compare the tradeoffs. A practical model shortlist starts with the work you need done.
Source snapshot Sep 25, 2026
01 / Start with the task
Different tests. Different strengths.
Highest observed means in six Epoch AI benchmark tables. These are source results, not Rubrex tests or universal winners. Coverage and evaluation settings vary.
01 / Long-horizon coding↗
Claude Fable 5.1 (high)
73.3%MirrorCode Mean score · SE not reported
Reimplement complete programs from their behavior and documentation.
Run started: .
Uses a large agent budget: up to seven days and 10 billion tokens per attempt. This is not a quick coding-assistant test.
In this snapshot, Claude Fable 5.1 (high) has the highest observed mean on MirrorCode; GPT-6 Astra (max) leads the included GPQA Diamond runs. Use the task-specific results below to select candidates, then test them on your own examples at a fixed budget.
02 / Explore the evidence
Compare within the same test.
MirrorCode
Reimplement complete programs from their behavior and documentation. Uses a large agent budget: up to seven days and 10 billion tokens per attempt. This is not a quick coding-assistant test.
Showing 8 of 8 matching evaluation runs. Ordered by mean score, highest first. ± is one standard error in percentage points, where supplied.
Scroll the table horizontally on smaller screens.
MirrorCode: Epoch AI internal evaluations
Model and setting
Mean score
Standard error
Run started
Claude Fable 5.1 (high)Anthropic
73.3%
Not reported
Sep 10, 2026
Claude Fable 5 (high)Anthropic
63.9%
± 10.4 pp
Aug 10, 2026
GPT-6 Astra (high)OpenAI
46.7%
Not reported
Aug 30, 2026
Claude Opus 4.7 (high)Anthropic
31.1%
± 9.0 pp
Aug 12, 2026
GPT-5.6 Sol (high)OpenAI
20.0%
± 9.2 pp
Aug 10, 2026
GPT-5.4 (high)OpenAI
15.6%
± 7.7 pp
Aug 10, 2026
GPT-5.5 (high)OpenAI
10.0%
± 6.0 pp
Aug 10, 2026
Gemini 3.1 Pro PreviewGoogle DeepMind
8.9%
± 4.6 pp
Aug 10, 2026
03 / Build your shortlist
Two models. Six different questions.
Compare model families using their highest observed mean score in each test. The exact setting appears below every result. This view does not equalize reasoning effort or compute budgets.
Claude Fable 5.1No resultMissing from this source snapshot
GPT-6 AstraNo resultMissing from this source snapshot
A missing score is unknown, not zero. Higher means can have overlapping uncertainty. These results are a shortlist signal, not proof of superiority on your workload.
04 / Make the budget explicit
What would the tokens cost?
A dated sample from OpenRouter’s public catalog. Prices are USD per million tokens at the listed base tier. The example uses 10,000 input tokens and 2,000 total billed output tokens for one request.
Scroll the table horizontally on smaller screens.
OpenRouter catalog prices, independent of the benchmark runs above
This is a token estimate, not cost per successful task. Extra reasoning tokens, retries, tools, search, media, routing/provider differences, and higher-context tiers can change the bill. Cache and batch discounts are excluded. Verify the linked model listing before purchase.
05 / Evidence you can trace
Know what the numbers mean.
One source for benchmark scores
215 model families appear across these six tables. We import Epoch AI’s internal-run mean scores, retain exact settings, and show supplied standard errors. We convert fractions to percentages and sort results. We do not average unrelated benchmarks into a new rating.
Sources retrieved Sep 25, 2026. This is a saved snapshot, not a live feed. The run date in each row can be older than the download date. New models enter different tests at different times; a missing result does not mean poor performance.
Reasoning effort, tools, scaffolds, and compute budgets affect results. A small mean-score gap is not a statistical significance test. Before switching models, evaluate task success, tail latency, failure severity, and total cost under your own operating conditions.
These benchmark tables do not measure API speed or user preference. Artificial Analysis provides performance comparisons; Arena provides preference-based comparisons. Their tests answer different questions, so their numbers are not merged into this table.
Benchmark data: Epoch AI, Capabilities & Benchmarking, used under CC BY 4.0. Selected, reformatted, and converted to percentages by Rubrex. Original benchmark creators and methods are linked in each category. No endorsement is implied. API price data: OpenRouter public model catalog, retrieved on the same date.
The source export can change after this snapshot. We publish no provider-sponsored ranking and make no claim that these scores establish production reliability.
Before you choose
A few useful answers.
Why is a newer model missing from some tests?
Independent evaluation takes time, and each benchmark covers a different set of models. We leave missing results empty. We do not infer a score from another model in the same family or from a launch claim.
Does the cheapest model offer the best value?
Only if it meets your quality bar. Compare cost per successful task, including retries and human review. A low token price can still produce an expensive workflow.
Can these results choose a model for RAG or customer support?
They can help build a shortlist. You still need representative documents, answerability cases, citation checks, tool boundaries, and tests of the complete application.