Ranking lists
The full ModelCap Index ranks every current language model by public benchmark evidence with published uncertainty. These lists cut that same board by one need at a time — open weights, context length, price, coding evidence, agent evidence, free access — and keep the Index position beside every row so nothing here contradicts the main rankings.
Snapshot as of 1 October 2026
The full board
Every current canonical language model with a public position, ordered by Index score, with evidence labels, prices and context.
Open-weight ranking
Language models whose weights can be downloaded — openly or behind a licence gate — ordered by their position on the live ModelCap Index.
MiMo-V2.6-Pro leads 173 models · ModelCap Index #6
Context window ranking
Ranked language models ordered by the largest published context window, with the ModelCap Index position, price and weight access beside each so a long context never hides a weak model.
Grok 4.20 Multi-Agent leads 122 models · ModelCap Index #13
Price ranking
Ranked language models with a listed API price, cheapest first by a 3:1 input-weighted blend of the lowest observed provider offer.
Ling 3.0 Flash VL leads 115 models · ModelCap Index #50
Coding ranking
Language models measured on at least two public coding boards (Arena Coding, SWE-bench bash-only, Terminal-Bench 2.1 with Terminus 2), ordered by the coding component of the ModelCap Index.
Kimi K3 leads 105 models · ModelCap Index #10
Agent ranking
Language models measured on at least two public agent boards (LMArena Agent, BFCL V4, WildClawBench), ordered by the agent component of the ModelCap Index.
GPT-5.5 leads 28 models · ModelCap Index #14
Free-tier ranking
Ranked language models that at least one API provider serves at no cost, ordered by ModelCap Index position.
Gemma 4 31B leads 11 models · ModelCap Index #28
Best open-weight LLMs; LLMs with the largest context windows; Cheapest LLM APIs; Best AI models for coding; Best AI models for agents; Free LLM APIs. Each is a filtered, re-ordered view of the same live ModelCap Index, never a separate score.
Best open-weight LLMs: MiMo-V2.6-Pro. Largest context window LLMs: Grok 4.20 Multi-Agent. Cheapest LLM APIs: Ling 3.0 Flash VL. Best AI models for coding: Kimi K3. Best AI models for agents: GPT-5.5. Free LLM APIs: Gemma 4 31B. As of 1 October 2026.
The main rankings page is the full ModelCap Index: every current canonical language model with a public position, ordered by Index score with published uncertainty. Each list here starts from that population, keeps the rows that match one published fact (open weights, a free tier, admitted coding or agent results…) and orders by one published figure (context, price, a component of the Index, or the Index rank itself).
Each list uses a published ModelCap snapshot. Publication and page updates happen separately; the snapshot date is shown under each heading.