LLM leaderboard
Ranked on published benchmark results and nothing else. A model with no published score is not last, it is untested, so it does not appear here.
| # | Model | Provider | Score | Evidence |
|---|---|---|---|---|
| 1 | MiniMax | 75.8 | 1 row · 20% of weighting | |
| 2 | 75.8 | 1 row · 20% of weighting | ||
| 3 | Anthropic | 75.6 | 1 row · 20% of weighting | |
| 4 | Anthropic | 74.4 | 1 row · 20% of weighting | |
| 5 | Z.ai | 72.8 | 1 row · 20% of weighting | |
| 6 | OpenAI | 72.8 | 1 row · 20% of weighting | |
| 7 | OpenAI | 72.8 | 1 row · 20% of weighting | |
| 8 | OpenAI | 71.8 | 1 row · 20% of weighting | |
| 9 | Moonshot AI | 70.8 | 1 row · 20% of weighting | |
| 10 | Anthropic | 70.8 | 1 row · 20% of weighting | |
| 11 | Anthropic | 70.6 | 1 row · 20% of weighting | |
| 12 | DeepSeek | 70.0 | 1 row · 20% of weighting | |
| 13 | Anthropic | 66.6 | 1 row · 20% of weighting | |
| 14 | OpenAI | 66.0 | 1 row · 20% of weighting | |
| 15 | OpenAI | 66.0 | 1 row · 20% of weighting | |
| 16 | OpenAI | 64.6 | 1 row · 20% of weighting | |
| 17 | Moonshot AI | 63.4 | 1 row · 20% of weighting | |
| 18 | MiniMax | 61.0 | 1 row · 20% of weighting | |
| 19 | Qwen | 60.4 | 1 row · 20% of weighting | |
| 20 | OpenAI | 59.8 | 1 row · 20% of weighting | |
| 21 | OpenAI | 58.4 | 1 row · 20% of weighting | |
| 22 | Z.ai | 55.4 | 1 row · 20% of weighting | |
| 23 | Z.ai | 54.2 | 1 row · 20% of weighting | |
| 24 | 53.6 | 1 row · 20% of weighting | ||
| 25 | OpenAI | 45.0 | 1 row · 20% of weighting |
The update feed
1–20 of 534
- Pricing
MiniMax M2.5 pricing changedModel: MiniMax M2.5 Input was $0.27 per 1M tokens, now $0.295 per 1M tokens Output was $0.95 per 1M tokens, now $1.2 per 1M tokens
- Pricing
GLM 4.5 Air pricing changedModel: GLM 4.5 Air Input was $0.2 per 1M tokens, now $0.13 per 1M tokens Output was $1.1 per 1M tokens, now $0.85 per 1M tokens
- Pricing
DeepSeek V4.1 Flash pricing changedModel: DeepSeek V4.1 Flash Input was $0.051 per 1M tokens, now $0.14 per 1M tokens Output was $0.567 per 1M tokens, now $0.42 per 1M tokens
- Pricing
DeepSeek Flash Latest pricing changedModel: DeepSeek Flash Latest Input was $0.02 per 1M tokens, now $0.036 per 1M tokens Output was $0.108 per 1M tokens, now $0.5 per 1M tokens
- Pricing
Kimi K3.1 Pro pricing changedModel: Kimi K3.1 Pro Input was $2 per 1M tokens, now $2.3 per 1M tokens Output was $10 per 1M tokens, now $11.5 per 1M tokens
- Pricing
gpt-oss-20b pricing changedModel: gpt-oss-20b Input was $0.03 per 1M tokens, now $0.075 per 1M tokens Output was $0.13 per 1M tokens, now $0.3 per 1M tokens
- Pricing
Qwen3.5-9B pricing changedModel: Qwen3.5-9B Input was $0.04 per 1M tokens, now $0.155 per 1M tokens Output was $0.059 per 1M tokens, now $0.228 per 1M tokens
- Pricing
Kimi K2.7 Code pricing changedModel: Kimi K2.7 Code Input was $0.6 per 1M tokens, now $0.7125 per 1M tokens Output was $3.5 per 1M tokens, now $3.5 per 1M tokens
- Pricing
DeepSeek Flash Latest pricing changedModel: DeepSeek Flash Latest Input was $0.038 per 1M tokens, now $0.02 per 1M tokens Output was $0.5 per 1M tokens, now $0.108 per 1M tokens
- Pricing
R1 0528 pricing changedModel: R1 0528 Input was $0.571 per 1M tokens, now $0.5 per 1M tokens Output was $2.286 per 1M tokens, now $2.15 per 1M tokens
- Pricing
Kimi K3.1 Pro pricing changedModel: Kimi K3.1 Pro Input was $0.2375 per 1M tokens, now $0.2232 per 1M tokens Output was $8.4114 per 1M tokens, now $8.4114 per 1M tokens
- Pricing
GLM 4.7 pricing changedModel: GLM 4.7 Input was $0.6 per 1M tokens, now $0.4 per 1M tokens Output was $2.2 per 1M tokens, now $1.75 per 1M tokens
- Pricing
DeepSeek V4 Flash 0731 pricing changedModel: DeepSeek V4 Flash 0731 Input was $0.0434 per 1M tokens, now $0.06 per 1M tokens Output was $0.3 per 1M tokens, now $0.18 per 1M tokens
- Pricing
Qwen3.5-9B pricing changedModel: Qwen3.5-9B Input was $0.143 per 1M tokens, now $0.039 per 1M tokens Output was $0.21 per 1M tokens, now $0.058 per 1M tokens
- Pricing
R1 0528 pricing changedModel: R1 0528 Input was $0.5 per 1M tokens, now $0.571 per 1M tokens Output was $2.15 per 1M tokens, now $2.286 per 1M tokens
- PricingParticle-1 pricing changed
Model: Particle-1 Input was $0.05 per 1M tokens, now $0.06 per 1M tokens Output was $0.5 per 1M tokens, now $0.5 per 1M tokens
- Pricing
Qwen3.5-9B pricing changedModel: Qwen3.5-9B Input was $0.167 per 1M tokens, now $0.156 per 1M tokens Output was $0.245 per 1M tokens, now $0.229 per 1M tokens
- PricingRipple 65B Preview pricing changed
Model: Ripple 65B Preview Input was $0.2 per 1M tokens, now $0.1 per 1M tokens Output was $0.6 per 1M tokens, now $0.3 per 1M tokens
- Pricing
DeepSeek V4 Pro 0423 pricing changedModel: DeepSeek V4 Pro 0423 Input was $0.95526 per 1M tokens, now $0.9 per 1M tokens Output was $1.91052 per 1M tokens, now $1.5 per 1M tokens
- Pricing
MiMo-V2.5-Pro pricing changedModel: MiMo-V2.5-Pro Input was $0.48 per 1M tokens, now $0.435 per 1M tokens Output was $1.8 per 1M tokens, now $0.87 per 1M tokens
Browse by brand
Every provider WriteWorks tracks. Each brand page carries its updates and its models, and every model has its own page underneath.
OpenAI67 models · 17 updates
Qwen55 models · 47 updates
Z.ai19 models · 111 updates
DeepSeek18 models · 193 updates
MiniMax9 models · 9 updates
Moonshot AI9 models · 79 updates- Particle AI1 model · 2 updates
Xiaomi5 models · 4 updates- Ricursive1 model · 1 update
Google33 models · 19 updates- Mancer1 model · 1 update
- Undi951 model · 1 update
Meta17 models · 13 updates
Anthropic21 models · 1 update
Mistral20 models · 6 updates
NVIDIA8 models · 12 updates- inclusionAI7 models · 3 updates
Tencent6 models · 5 updates- IBM2 models · 2 updates
- Reflection1 model · 3 updates
Upstage3 models · 1 update- Unbiased3 models · 1 update
Liquid AI3 models · 1 update
Thinking Machines3 models · 2 updates
xAI8 models
ByteDance7 models- AionLabs6 models
Amazon5 models
Perplexity5 models
Cohere4 models
Sakana AI4 models
Baidu3 models
Inception3 models- Nex AGI3 models
Nous Research3 models- Sao10K3 models
StepFun3 models- TheDrummer3 models
Inference.net2 models
Microsoft2 models- Moondream2 models
Morph2 models
Perceptron2 models
Poolside2 models- PrismML2 models
Reka2 models
Relace2 models- ScaleDown2 models
- Stealth2 models
- Anthracite-org1 model
- Apodex1 model
Arcee AI1 model- Dots Studio1 model
- Elorian1 model
Fireworks1 model- Gryphe1 model
- KT1 model
- Kwaipilot1 model
Meituan1 model- Motif Technologies1 model
- Recursive1 model
- Ricursiveai1 model
- TypeSafe1 model
Venice1 model
Writer1 model- XingChen1 model
LLM leaderboards, benchmarks and rankings, explained
What the ranking above does and does not tell you.
- What is an LLM leaderboard?
- An LLM leaderboard ranks large language models against a fixed set of benchmarks so their results can be read side by side. The ranking is only as good as the evidence under it: a score with no published source, or a model tested on three benchmarks sitting above one tested on thirty, tells you less than it appears to.
- How are these AI rankings calculated?
- Each model is scored per category from its published benchmark results, then the categories are weighted into one figure: agentic 22%, coding 20%, reasoning 17%, multimodal 12%, knowledge 12%, multilingual 7%, instruction following 5%, maths 5%. WriteWorks holds scores for 35 of 418 tracked models across 19 benchmarks. A category with no published evidence is left out and the rest re-weighted, never counted as zero.
- Which LLM is best?
- There is no single best model, and any leaderboard implying otherwise is answering a narrower question than it looks. The top of a general ranking is the model with the strongest average across weighted categories; the right model for a given job is usually the one leading the one category that job depends on, at a price the workload can carry.
- How often are the rankings and benchmarks updated?
- Model catalogues are re-read every three hours and provider pricing pages every six, so releases and price moves appear the same day. Benchmark scores are added as results are published, and every row on a model page carries the date it was confirmed.
- What is the difference between LLM benchmarks and an LLM leaderboard?
- A benchmark is one test with one score, such as SWE-bench for resolving real code issues. A leaderboard is the ranking that falls out of combining several and deciding what each is worth. The benchmarks are the evidence; the leaderboard is an argument about how to weigh it, which is why this one publishes its weights.
- Are these AI benchmarks independently verified?
- Each score is labelled. Verified means the score was read on the benchmark's own published results. Provider claim means the only source is the company that made the model. Both are shown, because dropping everything that cannot be verified independently would quietly favour the labs that publish least.
Are these models citing you, or your competitors?
WriteWorks tracks how often 10+ AI platforms, from ChatGPT and Claude to Gemini and Perplexity, mention your brand against your competitors, then helps you engineer the content that changes the answer.