AI Model Rankings 2026
The weekly-updated leaderboard of 42 AI models — full benchmark scores, prices, speed and context windows in one table.
Snapshot 2026-10-14 · updated weekly
This week's champions
AI model leaderboard — October 2026
Sorted by AMR aggregate score. Preview models are ranked separately and marked.
| Rank | Model | Vendor | AMR Score | GPQA | Context | Price (in / out) | Speed |
|---|---|---|---|---|---|---|---|
1— | Claude Opus 5.5 | Anthropic | 96.8 | 95.8% | 500K | $8 / $40 | 88 |
2▲1 | GPT-6.1 Sol | OpenAI | 95.9 | 94.9% | 1M | $6 / $36 | 112 |
3▼1 | GPT-6 Astra | OpenAI | 95.6 | 96% | 1M | $5 / $30 | 128 |
4— | Claude Fable 5.1 | Anthropic | 94.9 | 93.4% | 400K | $4 / $20 | 142 |
5— | Claude Sonnet 5.5 | Anthropic | 94.7 | 92.8% | 500K | $3 / $15 | 156 |
6— | Claude Fable 5 | Anthropic | 93.8 | 91.9% | 400K | $3.50 / $18 | 138 |
7▲2 | Gemini 4 Argon | 93.2 | 92.1% | 2M | $2.50 / $12 | 165 | |
8▼1 | Claude Opus 5 | Anthropic | 92.6 | 90.7% | 400K | $6 / $30 | 96 |
9— | GPT-6 Sol | OpenAI | 91.4 | 90.2% | 800K | $4 / $24 | 118 |
10NEW | Claude Mythos (Preview)Preview | Anthropic | 91.0 | 94.1% | 500K | $10 / $50 | 64 |
11NEW | Grok 5 PreviewPreview | xAI | 90.8 | 91.2% | 1M | $3 / $15 | 120 |
12— | GPT-5.6 Sol | OpenAI | 89.7 | 88.4% | 512K | $2.80 / $16 | 134 |
13— | Gemini 4.0 | 89.0 | 87.9% | 2M | $2.20 / $10 | 158 | |
14NEW | DeepSeek Harness | DeepSeek | 88.8 | 86.5% | 256K | $0.35 / $1.40 | 105 |
15▲1 | MiniMax H3 | MiniMax | 88.2 | 85.8% | 1M | $0.60 / $2.20 | 121 |
16NEW | Claude Haiku 5.5 | Anthropic | 87.9 | 83.2% | 500K | $0.80 / $4 | 232 |
17— | Kimi K3 | Moonshot AI | 87.5 | 93.5% | 512K | $0.50 / $2 | 98 |
18— | Muse Wonder Full | Meta | 86.8 | 88.1% | 512K | $0.90 / $3.20 | 88 |
19— | GLM-5.3 | Zhipu AI | 85.5 | 85.2% | 256K | $0.40 / $1.50 | 118 |
20— | Muse Wonder | Meta | 85.2 | 85.9% | 256K | $0.45 / $1.60 | 126 |
21— | DeepSeek V4.5 | DeepSeek | 84.5 | 83.7% | 256K | $0.30 / $1.20 | 112 |
22▲3 | GLM-5.3 Flash | Zhipu AI | 84.1 | 80.4% | 256K | $0.12 / $0.50 | 268 |
23▲1 | Qwen3.8 Next | Alibaba | 83.6 | 82.9% | 512K | $0.50 / $1.80 | 104 |
24— | Grok-4 Fast Reasoning | xAI | 83.0 | 81.5% | 2M | $0.80 / $3.20 | 210 |
25▲2 | Gemini 3.8 Flash | 82.7 | 78.9% | 1M | $0.15 / $0.60 | 342 | |
26— | Mistral Large 3.5 | Mistral AI | 82.3 | 80.7% | 256K | $0.80 / $2.40 | 132 |
27— | Gemini 3 Pro | 81.0 | 80.9% | 1M | $1.50 / $6 | 168 | |
28— | Gemini 3.1 Pro | 80.2 | 79.8% | 1M | $1.25 / $5 | 172 | |
29— | Qwen3.8 27B | Alibaba | 79.8 | 76.8% | 256K | $0.15 / $0.60 | 186 |
30— | MiniMax H3 Flash | MiniMax | 78.8 | 72.6% | 512K | $0.10 / $0.40 | 358 |
31— | GPT-5.6 Luna | OpenAI | 78.4 | 74.2% | 512K | $1.20 / $5.50 | 286 |
32— | MiMo-V2.6-Pro | Xiaomi | 78.0 | 75.4% | 256K | $0.25 / $1 | 142 |
33— | Kimi K3 Flash | Moonshot AI | 76.0 | 70.2% | 256K | $0.14 / $0.55 | 296 |
34— | Qwen3.8 Flash | Alibaba | 74.5 | 66.8% | 256K | $0.06 / $0.25 | 412 |
35— | Gemma 4 27B | 72.0 | 67.4% | 256K | $0.18 / $0.70 | 174 | |
36NEW | Strata 2 | Strata Labs | 70.5 | 62.8% | 128K | $0.12 / $0.45 | 248 |
37NEW | Jev 1 | Jev Labs | 68.9 | 59.4% | 128K | $0.08 / $0.30 | 322 |
38NEW | Clef 2 | Clef AI | 67.2 | 56.9% | 128K | $0.10 / $0.35 | 276 |
39— | Mercury 2.5 | Inception Labs | 62.0 | 48.2% | 128K | $0.10 / $0.40 | 583 |
40— | Celeris-1 | Celeris | 58.2 | 42.6% | 128K | $0.20 / $0.80 | 1431 |
41NEW | Laya 1 | Laya Research | 55.4 | 38.4% | 64K | $0.05 / $0.20 | 368 |
42NEW | Pixel Canary | 48.0 | 31.2% | 32K | Free / Free | 210 |
Compare any two AI models
Pick two models and get a side-by-side breakdown across every benchmark, price, speed and context window — then copy the result as an image for social media.
CompareAI model rankings FAQ
What is the best AI model right now?
Claude Opus 5.5 is the #1 AI model as of October 2026, with an AMR aggregate score of 96.8. It leads SWE-Bench Verified (81.4%) and ranks top-two on GPQA Diamond and AIME 2025. GPT-6.1 Sol and GPT-6 Astra follow within one aggregate point.
How often are these AI model rankings updated?
The leaderboard is refreshed weekly, and new model pages go live within 48 hours of a release. Benchmark scores are re-verified against public leaderboards and vendor rate cards; each row carries the snapshot date so you always know how fresh the data is.
What is the cheapest good AI model?
Gemini 3.1 Pro is the value pick among closed models at $1.25 in / $5.00 out per million tokens. In the open tier, DeepSeek Harness ($0.35 / $1.40) offers the best quality per dollar of any model tracked, scoring within nine points of the #1 model.
Which AI model has the biggest context window?
Gemini 4 Argon and Gemini 4.0 both offer a 2M token context window, tied with Grok-4 Fast Reasoning. That is four times the 500K window of Claude Opus 5.5 and enough to fit several large codebases in a single prompt.
What is the fastest AI model?
Celeris-1 is the fastest model tracked at 1,431 output tokens per second, built for autocomplete and real-time agents. Among models with near-frontier quality, MiniMax H3 Flash (358 tok/s) and Gemini 3.8 Flash (342 tok/s) lead the field.
What is the best open-source AI model?
Kimi K3 is the best open-weights AI model, ranking #15 overall with a 93.5% GPQA Diamond score — the highest of any model you can self-host. MiniMax H3 and DeepSeek Harness follow closely, both with 1M-class context windows.