robotmunki.commodel-selector

LLM Model Selector

Top Models — Default View

Top 8 by capability score — apply levers to filter by task, cost, context, or vendor

Nov 2026 data
1

Gemini 4 Carbon

Google
AA 59

AA v5.0 #1 (59, high) since Oct 28 — but limited to vetted cyber defenders via Google's Fairwind program; no public API date.

$20/M out1M ctx
AA v5 #1RestrictedCyber-vetted only
2

Claude Opus 5.5

Anthropic
AA 58

Highest-scoring generally-available model on AA v5.0 (58, max); leads FrontierCode 2.0, HLE-Diamond and AA-Omniscience; cheaper than Opus 5 at $4/$20.

$20/M out1M ctx
AA v5 58Best availableFrontierCode leader
3

GPT-6 Astra

OpenAI
AA 57

OpenAI's flagship (Sep 3) — leads Terminal-Bench-Science and CritPt v2; reportedly access-gated on cyber grounds.

$50/M out1.05M ctx
AA v5 57Science leaderPremium
4

GPT-6.1 Sol

OpenAI
AA 55

Best cost-efficiency near the top — 55 on AA v5.0 for ~$1.27 per index task (vs ~$10 for Opus 5.5 max).

$10/M out1M ctx
AA v5 55Best value at the topToken-efficient
5

Claude Sonnet 5.5

Anthropic
AA 54

Leads AA-Briefcase knowledge work (1,812 Elo) at $2/$10 — but uses ~193K output tokens per index task at max effort.

$10/M out1M ctx
AA v5 54Knowledge work leaderToken-hungry
6

Claude Mythos 5.1

Anthropic
—

Restricted to US orgs in Anthropic's Cyber and Life Sciences verification programs.

$50/M out1M ctx
RestrictedVerification programs
7

GPT-6.1 Terra

OpenAI
AA 51

$1/$5 mid-tier that still scores 51 on v5.0.

$5/M out1M ctx
AA v5 51Mid-tierCheap
8

Gemini 4 Argon

Google
AA 50

Google's generally-available Gemini 4 (GA Oct 13) — leads AutomationBench-AA, 1M output-token limit; intro $2/$10 rises to $4/$20 on Dec 1.

$10/M out1M ctx
AA v5 501M outputIntro $2/$10 → $4/$20 Dec 1

Lever Reference

1
Capability

AA Intelligence Index v5.0 (knowledge work, coding, science, general — ~60% private evals)

2
Task Fit

Peak AA score does not equal best task fit (e.g. coding)

3
Context Window

Standard ≤262K · Large 263K–999K · Massive 1M–10M

4
Cost Per Token

Budget ≤ $4/M · Mid $4–15/M · Premium $15+/M output

5
Deployment

API-only vs open-weight self-host vs both

6
Vendor Origin

Geographic / residency-style grouping for vendor choice