DOAClaude Opus 5.5
100.0/100 · DOA rankingRank 1 · Arena (text) 1,503.7 · Epoch AI (ECI) 167.3
Anthropic
- Pro · US$20/month
- Max · from US$100/month
Source: claude.com (opens in a new tab) · 26 September 2026
28 models · 11 apps
AI models
DOA ranking = normalised average of external sources (Arena and Epoch AI). DOA adds no opinion. Below, which app and plan you can use each model in.
DOA ranking
Simple average of Arena and Epoch AI, each normalised from 0 to 100 across the models on this list that appear in both; only computed with 2 or more sources. It is a relative scale (100 = the best on the list in that source, 0 = the lowest), not an absolute mark. In brackets, the normalised figure that goes into the average. Arena data from 2026-10-02 and Epoch AI data from 2026-10-08, downloaded 2026-10-08.
| Rank | Model | DOA ranking (average) | Arena (text) | Epoch AI (ECI) |
|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropic | 100.0 | 1,503.7 (opens in a new tab) (100.0)claude-opus-5.5-high · 2026-10-02 | 167.3 (opens in a new tab) (100.0)Claude Opus 5.5 · 2026-10-08 |
| 2 | Claude Fable 5.1Anthropic | 94.9 | 1,501.1 (opens in a new tab) (98.6)claude-fable-5.1-max · 2026-10-02 | 164.7 (opens in a new tab) (91.1)Claude Fable 5.1 · 2026-10-08 |
| 3 | GPT-6.1 SolOpenAI | 92.4 | 1,483.2 (opens in a new tab) (89.0)gpt-6.1-sol-max · 2026-10-02 | 166.1 (opens in a new tab) (95.8)GPT-6.1 Sol · 2026-10-08 |
| 4 | GPT-6 AstraOpenAI | 91.4 | 1,477.1 (opens in a new tab) (85.7)gpt-6-astra-max · 2026-10-02 | 166.5 (opens in a new tab) (97.0)GPT-6 Astra · 2026-10-08 |
| 5 | Claude Sonnet 5.5Anthropic | 87.3 | 1,471.0 (opens in a new tab) (82.4)claude-sonnet-5.5-xhigh · 2026-10-02 | 165.0 (opens in a new tab) (92.2)Claude Sonnet 5.5 · 2026-10-08 |
| 6 | Gemini 3.8 FlashGoogle | 79.6 | 1,494.8 (opens in a new tab) (95.2)gemini-3.8-flash-high · 2026-10-02 | 156.7 (opens in a new tab) (64.0)Gemini 3.8 Flash · 2026-10-08 |
| 7 | Muse Spark 1.3Meta | 79.5 | 1,494.3 (opens in a new tab) (94.9)muse-spark-1.3-max · 2026-10-02 | 156.8 (opens in a new tab) (64.1)Muse Spark 1.3 · 2026-10-08 |
| 8 | Kimi K3Moonshot AI (Kimi) | 79.1 | 1,488.4 (opens in a new tab) (91.8)kimi-k3-max · 2026-10-02 | 157.5 (opens in a new tab) (66.5)Kimi K3 · 2026-10-08 |
| 9 | Gemini 3.1 Pro (preview)Google | 74.2 | 1,487.0 (opens in a new tab) (91.0)gemini-3.1-pro-preview · 2026-10-02 | 154.8 (opens in a new tab) (57.4)Gemini 3.1 Pro · 2026-10-08 |
| 10 | GLM-5.3Z.ai | 73.4 | 1,478.5 (opens in a new tab) (86.5)glm-5.3-max · 2026-10-02 | 155.6 (opens in a new tab) (60.3)GLM-5.3 · 2026-10-08 |
| 11 | Qwen3.8 MaxAlibaba | 73.3 | 1,481.6 (opens in a new tab) (88.1)qwen3.8-max · 2026-10-02 | 155.1 (opens in a new tab) (58.4)Qwen3.8 Max (0902) · 2026-10-08 |
| 12 | DeepSeek V4.1 FlashDeepSeek | 71.1 | 1,474.4 (opens in a new tab) (84.3)deepseek-v4.1-flash-max · 2026-10-02 | 154.9 (opens in a new tab) (57.9)DeepSeek V4.1 Flash · 2026-10-08 |
| 13 | DeepSeek V4 ProDeepSeek | 69.0 | 1,464.3 (opens in a new tab) (78.8)deepseek-v4-pro-high-20260813 · 2026-10-02 | 155.3 (opens in a new tab) (59.3)DeepSeek V4 Pro 0813 · 2026-10-08 |
| 14 | GPT-6 LunaOpenAI | 64.9 | 1,442.9 (opens in a new tab) (67.3)gpt-6-luna-max · 2026-10-02 | 156.3 (opens in a new tab) (62.6)GPT-6 Luna · 2026-10-08 |
| 15 | Grok 4.7SpaceXAI | 60.2 | 1,442.5 (opens in a new tab) (67.1)grok-4.7-xhigh · 2026-10-02 | 153.5 (opens in a new tab) (53.2)Grok 4.7 · 2026-10-08 |
| 16 | Gemini 3.5 Flash-LiteGoogle | 49.3 | 1,455.0 (opens in a new tab) (73.8)gemini-3.5-flash-lite · 2026-10-02 | 145.1 (opens in a new tab) (24.8)Gemini 3.5 Flash-Lite · 2026-10-08 |
| 17 | MiniMax-M3MiniMax | 48.4 | 1,440.1 (opens in a new tab) (65.8)minimax-m3 · 2026-10-02 | 147.0 (opens in a new tab) (30.9)MiniMax-M3 · 2026-10-08 |
| 18 | Gemma 4 31BGoogle | 44.7 | 1,452.8 (opens in a new tab) (72.7)gemma-4-31b · 2026-10-02 | 142.7 (opens in a new tab) (16.7)Gemma 4 31B IT · 2026-10-08 |
| 19 | Mistral Medium 3.5Mistral AI | 35.4 | 1,426.9 (opens in a new tab) (58.8)mistral-medium-3.5 · 2026-10-02 | 141.4 (opens in a new tab) (12.0)Mistral Medium 3.5 · 2026-10-08 |
| 20 | Claude Haiku 4.5Anthropic | 33.6 | 1,413.8 (opens in a new tab) (51.7)claude-haiku-4-5-20251001 · 2026-10-02 | 142.4 (opens in a new tab) (15.6)Claude Haiku 4.5 · 2026-10-08 |
| 21 | gpt-oss-120bOpenAI | 12.7 | 1,351.6 (opens in a new tab) (18.3)gpt-oss-120b · 2026-10-02 | 139.9 (opens in a new tab) (7.2)gpt-oss-120b · 2026-10-08 |
| 22 | gpt-oss-20bOpenAI | 0.0 | 1,317.6 (opens in a new tab) (0.0)gpt-oss-20b · 2026-10-02 | 137.8 (opens in a new tab) (0.0)gpt-oss-20b · 2026-10-08 |
| — | MiMo-V2.6-ProXiaomi | No ranking: 1 source | 1,480.1 (opens in a new tab)mimo-v2.6-pro · 2026-10-02 | not listed |
| — | Mistral Large 3Mistral AI | No ranking: 1 source | 1,413.9 (opens in a new tab)mistral-large-3 · 2026-10-02 | not listed |
| — | Mistral Small 4Mistral AI | Not in the sources | not listed | not listed |
| — | Muse GlimmerMeta | No ranking: 1 source | 1,425.0 (opens in a new tab)muse-glimmer · 2026-10-02 | not listed |
| — | Qwen3.8 27BAlibaba | No ranking: 1 source | 1,437.7 (opens in a new tab)qwen3.8-27b · 2026-10-02 | not listed |
| — | Qwen3.8 Flash NextAlibaba | Not in the sources | not listed | not listed |
Models with fewer than 2 sources come last, in alphabetical order: we show their original figures but no DOA ranking, because one source alone gives no average.
Sources and licences
- Arena: dataset lmarena-ai/leaderboard-dataset (opens in a new tab) on Hugging Face: text_style_control configuration (style control, Arena's default view), latest split, overall category, rating column. Licence CC BY 4.0 (opens in a new tab) · data from 2026-10-02, downloaded 2026-10-08.
- Epoch AI: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Only the Epoch Capabilities Index (eci_scores.csv) is used. epoch.ai/benchmarks/eci (opens in a new tab) · benchmark_data.zip (opens in a new tab). Licence CC BY 4.0 (opens in a new tab) · data from 2026-10-08, downloaded 2026-10-08.
Changes by DOA: we normalised (0-100) and averaged these figures, so the DOA ranking is not a figure from any of the sources. Arena and Epoch AI take no part in the DOA ranking and do not endorse it; their names appear only as attribution.
Where to use each model
When the official source does not make clear which plans include a model, we mark it «Check» with a link. Prices change: always confirm on the official site.
28 of 28 models
Claude Fable 5.1
94.9/100 · DOA rankingRank 2 · Arena (text) 1,501.1 · Epoch AI (ECI) 164.7
Anthropic
- Pro · US$20/month (with usage credits)
- Max · from US$100/month (50% of weekly limits)
Source: claude.com (opens in a new tab) · 30 September 2026
GPT-6.1 Sol
92.4/100 · DOA rankingRank 3 · Arena (text) 1,483.2 · Epoch AI (ECI) 166.1
OpenAI
ChatGPT ↗ (opens in a new tab)
- Plus · €23/month (limited)
- Pro · from €103/month (expanded)
replaces GPT-6 Sol since 29-09-2026
Source: chatgpt.com (opens in a new tab) · 30 September 2026
GPT-6 Astra
91.4/100 · DOA rankingRank 4 · Arena (text) 1,477.1 · Epoch AI (ECI) 166.5
OpenAI
ChatGPT ↗ (opens in a new tab)
- Plus · €23/month (limited)
- Pro · from €103/month (expanded)
Source: chatgpt.com (opens in a new tab) · 26 September 2026
Claude Sonnet 5.5
87.3/100 · DOA rankingRank 5 · Arena (text) 1,471.0 · Epoch AI (ECI) 165.0
Anthropic
- Free
- Pro · US$20/month
- Max · from US$100/month
Source: claude.com (opens in a new tab) · 30 September 2026
Gemini 3.8 Flash
79.6/100 · DOA rankingRank 6 · Arena (text) 1,494.8 · Epoch AI (ECI) 156.7
Plans: check the official site ↗ (opens in a new tab)
The Spain subscriptions page details Gemini 3.1 Pro, not this Flash model per plan.
Source: gemini.google (opens in a new tab) · 27 September 2026
Muse Spark 1.3
79.5/100 · DOA rankingRank 7 · Arena (text) 1,494.3 · Epoch AI (ECI) 156.8
Meta
Meta AI ↗ (opens in a new tab)
- Free (everyday use free)
Meta confirms Meta AI stays free. It does not publish which Muse model sits on Core vs Premium.
Source: about.fb.com (opens in a new tab) · 27 September 2026
Kimi K3
79.1/100 · DOA rankingRank 8 · Arena (text) 1,488.4 · Epoch AI (ECI) 157.5
Moonshot AI (Kimi) Open weights (downloadable)
Plans: check the official site ↗ (opens in a new tab)
Source: kimi.com (opens in a new tab) · 26 September 2026
Gemini 3.1 Pro (preview)
74.2/100 · DOA rankingRank 9 · Arena (text) 1,487.0 · Epoch AI (ECI) 154.8
- Free (variable access)
- Google AI Plus · €4.99/month (includes the free plan)
- Google AI Pro · €21.99/month (higher limits)
- Google AI Ultra · from €99.99/month (higher limits)
Source: gemini.google (opens in a new tab) · 26 September 2026
GLM-5.3
73.4/100 · DOA rankingRank 10 · Arena (text) 1,478.5 · Epoch AI (ECI) 155.6
Z.ai Open weights (downloadable)
Qwen3.8 Max
73.3/100 · DOA rankingRank 11 · Arena (text) 1,481.6 · Epoch AI (ECI) 155.1
Alibaba
Qwen Chat ↗ (opens in a new tab)
- Free (consumer chat, no Plus table)
The chat app does not publish model-per-paid-plan because there is no consumer plan table.
Source: chat.qwen.ai (opens in a new tab) · 27 September 2026
DeepSeek V4.1 Flash
71.1/100 · DOA rankingRank 12 · Arena (text) 1,474.4 · Epoch AI (ECI) 154.9
DeepSeek Open weights (downloadable)
DeepSeek ↗ (opens in a new tab)
- Free (chat has no subscription)
Official chat does not sell plans; Flash is also billed separately on the API.
Source: chat.deepseek.com (opens in a new tab) · 27 September 2026
DeepSeek V4 Pro
69.0/100 · DOA rankingRank 13 · Arena (text) 1,464.3 · Epoch AI (ECI) 155.3
DeepSeek Open weights (downloadable)
DeepSeek ↗ (opens in a new tab)
- Free (chat has no subscription)
Official chat does not sell plans; V4 Pro is also billed separately on the API.
Source: chat.deepseek.com (opens in a new tab) · 27 September 2026
GPT-6 Luna
64.9/100 · DOA rankingRank 14 · Arena (text) 1,442.9 · Epoch AI (ECI) 156.3
OpenAI
ChatGPT ↗ (opens in a new tab)
- Free
- Go · €8/month
- Plus · €23/month (limited)
- Pro · from €103/month (expanded)
Source: chatgpt.com (opens in a new tab) · 9 October 2026
Grok 4.7
60.2/100 · DOA rankingRank 15 · Arena (text) 1,442.5 · Epoch AI (ECI) 153.5
SpaceXAI
Plans: check the official site ↗ (opens in a new tab)
x.ai/pricing (14 Oct 2026) mentions “the latest Grok models” on Premium without naming a version: no plan until the official page names it.
Source: x.ai (opens in a new tab) · 14 October 2026
Gemini 3.5 Flash-Lite
49.3/100 · DOA rankingRank 16 · Arena (text) 1,455.0 · Epoch AI (ECI) 145.1
Plans: check the official site ↗ (opens in a new tab)
The Spain subscriptions page details Gemini 3.1 Pro, not this Flash model per plan.
Source: gemini.google (opens in a new tab) · 27 September 2026
MiniMax-M3
48.4/100 · DOA rankingRank 17 · Arena (text) 1,440.1 · Epoch AI (ECI) 147.0
MiniMax Open weights (downloadable)
MiniMax Agent ↗ (opens in a new tab)
Plans: check the official site ↗ (opens in a new tab)
Shown in the app's model picker.
Source: agent.minimax.io (opens in a new tab) · 26 September 2026
Gemma 4 31B
44.7/100 · DOA rankingRank 18 · Arena (text) 1,452.8 · Epoch AI (ECI) 142.7
Google Open weights (downloadable)
No confirmed official consumer app.
Mistral Medium 3.5
35.4/100 · DOA rankingRank 19 · Arena (text) 1,426.9 · Epoch AI (ECI) 141.4
Mistral AI Open weights (downloadable)
Claude Haiku 4.5
33.6/100 · DOA rankingRank 20 · Arena (text) 1,413.8 · Epoch AI (ECI) 142.4
Anthropic
- Free
- Pro · US$20/month
- Max · from US$100/month
Source: claude.com (opens in a new tab) · 26 September 2026
gpt-oss-120b
12.7/100 · DOA rankingRank 21 · Arena (text) 1,351.6 · Epoch AI (ECI) 139.9
OpenAI Open weights (downloadable)
No confirmed official consumer app.
gpt-oss-20b
0.0/100 · DOA rankingRank 22 · Arena (text) 1,317.6 · Epoch AI (ECI) 137.8
OpenAI Open weights (downloadable)
No confirmed official consumer app.
MiMo-V2.6-Pro
No ranking: 1 source · Arena (text) 1,480.1
Xiaomi Open weights (downloadable)
No confirmed official consumer app.
Mistral Large 3
No ranking: 1 source · Arena (text) 1,413.9
Mistral AI Open weights (downloadable)
No confirmed official consumer app.
Mistral Small 4
Not in the sources
Mistral AI Open weights (downloadable)
Muse Glimmer
No ranking: 1 source · Arena (text) 1,425.0
Meta Open weights (downloadable)
Meta AI ↗ (opens in a new tab)
- Free (everyday use free)
Meta confirms Meta AI stays free. It does not publish which Muse model sits on Core vs Premium.
Source: about.fb.com (opens in a new tab) · 27 September 2026
Qwen3.8 27B
No ranking: 1 source · Arena (text) 1,437.7
Alibaba Open weights (downloadable)
Qwen Chat ↗ (opens in a new tab)
- Free (consumer chat, no Plus table)
The chat app does not publish model-per-paid-plan because there is no consumer plan table.
Source: chat.qwen.ai (opens in a new tab) · 27 September 2026
Qwen3.8 Flash Next
Not in the sources
Alibaba Open weights (downloadable)
Qwen Chat ↗ (opens in a new tab)
- Free (consumer chat, no Plus table)
The chat app does not publish model-per-paid-plan because there is no consumer plan table.
Source: chat.qwen.ai (opens in a new tab) · 27 September 2026
App prices
Monthly price of each plan according to the official pricing page (checked 30 September 2026). Euros only when the page gives the price for Spain; otherwise the currency it shows, labelled.
The table is wider than the screen: swipe it sideways →
| App | Plan | Price per month | Source |
|---|---|---|---|
| ChatGPTOfficial chatgpt.com/pricing (checked 2026-10-14 from Spain): €/month. Pro comes in three usage tiers; the help centre lists US$100, US$200 and US$500/month. | Free | Free | chatgpt.com (opens in a new tab) |
| Go | €8/month | ||
| Plus | €23/month | ||
| Pro | from €103/month | ||
| ClaudeOfficial claude.com/pricing: US$ excluding tax. Pro US$20/month or US$17/month billed annually (US$200). Max from US$100/month. No euro price on that page. | Free | Free | claude.com (opens in a new tab) |
| Pro | US$20/month | ||
| Max | from US$100/month | ||
| DeepSeekchat.deepseek.com has no paid subscription. Paid pricing is the token API (official Models & Pricing). Consumer app: free plan. | Free | Free | api-docs.deepseek.com (opens in a new tab) |
| GeminiOnly app here with official Spain prices: gemini.google/es/subscriptions. Plus €4.99, Pro €21.99, Ultra from €99.99/month (another tier €219.99). | Free | Free | gemini.google (opens in a new tab) |
| Google AI Plus | €4.99/month | ||
| Google AI Pro | €21.99/month | ||
| Google AI Ultra | from €99.99/month | ||
| Grokx.ai/pricing (2026-10-14) lists one subscription for Grok and X: Plus US$8, Premium US$30, Super US$100 and Ultra US$200 a month, excluding tax, plus a free plan. | Free | Free | x.ai (opens in a new tab) |
| Plus | US$8/month | ||
| Premium | US$30/month | ||
| Super | US$100/month | ||
| Ultra | US$200/month | ||
| KimiPrices in US$, monthly billing (cheaper billed annually). | Free | Free | kimi.com (opens in a new tab) |
| Plus | US$19/month | ||
| Pro | US$39/month | ||
| Max | US$99/month | ||
| Ultra | US$199/month | ||
| Meta AIMeta says everyday Meta AI stays free. Meta One (official 15 Sep 2026) adds extra usage from US$7.99 (Core) and US$19.99 (Premium). The note does not map each Muse model to a plan. | Free | Free | about.fb.com (opens in a new tab) |
| Meta One Core | US$7.99/month | ||
| Meta One Premium | US$19.99/month | ||
| MiniMax Agent | Check ↗ (opens in a new tab) · agent.minimax.io/pricing does not show prices without signing in. Left as consult. | agent.minimax.io (opens in a new tab) | |
| Mistral VibePrices in US$ excluding tax, as shown on the official page when we checked. | Free | Free | mistral.ai (opens in a new tab) |
| Pro | US$14.99/month | ||
| Qwen Chatchat.qwen.ai does not publish a Plus-style table. Consumer chat is presented as free; paid API / Qwen Work products are separate. | Free | Free | chat.qwen.ai (opens in a new tab) |
| Z.ai Chat | Check ↗ (opens in a new tab) · chat.z.ai does not show a public plan table without an account. Left as consult. | chat.z.ai (opens in a new tab) | |