← Back to home
Google
Reference
LLM Overview
Compare models by provider, pricing, context window, and strengths.
🏆 Best Overall
GPT-5.5 or Claude Opus 4.8 for complex tasks requiring top-tier reasoning.
💰 Best Value
DeepSeek V4 Flash ($0.18/M) or Qwen3 235B ($0.10/M) — near-frontier quality.
💻 Best for Coding
Qwen3 Coder or Claude Sonnet 4.6 for production code.
🔓 Best Open Source
Llama 4 Maverick or DeepSeek V4 Pro — open weights, self-hostable.
📏 Longest Context
Llama 4 Scout (10M tokens) or Grok 4.20 (2M tokens).
🤖 Best for Agents
GLM-5.2 — designed for agentic workflows with strong tool-use.
| Model | Provider | Context | Input $/M | Output $/M | License | Best For |
|---|---|---|---|---|---|---|
| Claude Haiku 4.5 | Anthropic | 200K tokens | $1.00 | $5.00 | Proprietary | FastAffordableProduction-ready |
| Claude Sonnet 4.6 | Anthropic | 1M tokens | $3.00 | $15.00 | Proprietary | Balanced quality/costCodeAnalysis |
| Claude Opus 4.8 | Anthropic | 1M tokens | $5.00 | $25.00 | Proprietary | Long contextWriting qualityComplex reasoning |
| Claude Opus 4.8 Fast | Anthropic | 1M tokens | $10.00 | $50.00 | Proprietary | Fast reasoningHigh quality |
| Command A | Cohere | 256K tokens | $2.50 | $10.00 | Proprietary | RAGEnterpriseRetrieval |
| DeepSeek V4 Flash | DeepSeek | 1M tokens | $0.09 | $0.18 | Open | Ultra-cheapLong contextFast |
| DeepSeek V4 Pro | DeepSeek | 1M tokens | $0.43 | $0.87 | Open | Excellent valueReasoningCode |
| DeepSeek R1 | DeepSeek | 164K tokens | $0.50 | $2.15 | Open | ReasoningMathScience |
| Gemma 4 31B | 262K tokens | $0.12 | $0.35 | Open | Open weightsSelf-hostableGood quality | |
| Gemini 2.5 Flash Lite | 1M tokens | $0.10 | $0.40 | Proprietary | Ultra-cheap1M context | |
| Gemini 2.5 Flash | 1M tokens | $0.30 | $2.50 | Proprietary | Very cheapFastVolume workloads | |
| Gemini 3.5 Flash | 1M tokens | $1.50 | $9.00 | Proprietary | FastMultimodalGood value | |
| Gemini 3.1 Pro | 1M tokens | $2.00 | $12.00 | Proprietary | Long contextMultimodalReasoning | |
| Llama 4 Scout | Meta | 10M tokens | $0.10 | $0.30 | Open | 10M contextOpen weightsCheap |
| Llama 3.3 70B | Meta | 131K tokens | $0.10 | $0.32 | Open | Solid all-rounderOpen weightsSelf-hostable |
| Llama 4 Maverick | Meta | 1M tokens | $0.15 | $0.60 | Open | Open weightsLong contextMultimodal |
| Mistral Small 3.2 | Mistral | 128K tokens | $0.07 | $0.20 | Open | CheapFastGood quality |
| Codestral 2508 | Mistral | 256K tokens | $0.30 | $0.90 | Open | Code generationFast |
| Mistral Large 2512 | Mistral | 262K tokens | $0.50 | $1.50 | Open | EuropeanGood multilingualCode |
| Nemotron 3 Super 120B | NVIDIA | 1M tokens | $0.09 | $0.45 | Open | Good valueLong context |
| Nemotron 3 Ultra 550B | NVIDIA | 1M tokens | $0.50 | $2.20 | Open | EnterpriseLong context |
| GPT-OSS 120B | OpenAI | 131K tokens | $0.04 | $0.18 | Open | Open weightsSelf-hostableGood value |
| GPT-5 Nano | OpenAI | 400K tokens | $0.05 | $0.40 | Proprietary | Ultra-cheapFast inference |
| GPT-5 Mini | OpenAI | 400K tokens | $0.25 | $2.00 | Proprietary | FastAffordableGood for simple tasks |
| o4-mini | OpenAI | 200K tokens | $1.10 | $4.40 | Proprietary | ReasoningMathScience |
| GPT-5.2 | OpenAI | 400K tokens | $1.75 | $14.00 | Proprietary | Balanced performance/costCode generationTool use |
| GPT-5.4 | OpenAI | 1M tokens | $2.50 | $15.00 | Proprietary | Long contextReasoningMultimodal |
| GPT-5.5 | OpenAI | 1M tokens | $5.00 | $30.00 | Proprietary | General reasoningCode generationMultimodal |
| Qwen3 235B | Qwen (Alibaba) | 262K tokens | $0.09 | $0.10 | Open | Excellent valueReasoning |
| Qwen3 Coder | Qwen (Alibaba) | 1M tokens | $0.22 | $1.80 | Open | Code generation1M contextOpen weights |
| Qwen3.5 397B | Qwen (Alibaba) | 256K tokens | $0.39 | $2.45 | Open | Large MoEReasoningOpen weights |
| Qwen3.6 Max | Qwen (Alibaba) | 262K tokens | $1.04 | $6.24 | Proprietary | MultilingualStrong reasoning |
| Grok 4.20 | xAI | 2M tokens | $1.25 | $2.50 | Proprietary | Real-time info2M contextLess filtered |
| Grok 4.3 | xAI | 1M tokens | $1.25 | $2.50 | Proprietary | ReasoningReal-time data |
| GLM-4.7 Flash | Z-AI (Zhipu) | 200K tokens | $0.06 | $0.40 | Open | Very cheapFast |
| GLM-5 | Z-AI (Zhipu) | 200K tokens | $0.60 | $1.92 | Open | Good valueMultilingual |
| GLM-5.2 | Z-AI (Zhipu) | 1M tokens | $0.95 | $3.00 | Open | Agentic tasksTool useMultilingual |
By Provider
OpenAI
7 models · from $0.18/M
GPT-OSS 120B$0.18
GPT-5 Nano$0.40
GPT-5 Mini$2.00
o4-mini$4.40
GPT-5.2$14.00
GPT-5.4$15.00
GPT-5.5$30.00
Anthropic
4 models · from $5.00/M
Claude Haiku 4.5$5.00
Claude Sonnet 4.6$15.00
Claude Opus 4.8$25.00
Claude Opus 4.8 Fast$50.00
5 models · from $0.35/M
Gemma 4 31B$0.35
Gemini 2.5 Flash Lite$0.40
Gemini 2.5 Flash$2.50
Gemini 3.5 Flash$9.00
Gemini 3.1 Pro$12.00
xAI
2 models · from $2.50/M
Grok 4.20$2.50
Grok 4.3$2.50
DeepSeek
3 models · from $0.18/M
DeepSeek V4 Flash$0.18
DeepSeek V4 Pro$0.87
DeepSeek R1$2.15
Z-AI (Zhipu)
3 models · from $0.40/M
GLM-4.7 Flash$0.40
GLM-5$1.92
GLM-5.2$3.00
Meta
3 models · from $0.30/M
Llama 4 Scout$0.30
Llama 3.3 70B$0.32
Llama 4 Maverick$0.60
Mistral
3 models · from $0.20/M
Mistral Small 3.2$0.20
Codestral 2508$0.90
Mistral Large 2512$1.50
Qwen (Alibaba)
4 models · from $0.10/M
Qwen3 235B$0.10
Qwen3 Coder$1.80
Qwen3.5 397B$2.45
Qwen3.6 Max$6.24
NVIDIA
2 models · from $0.45/M
Nemotron 3 Super 120B$0.45
Nemotron 3 Ultra 550B$2.20
Cohere
1 model · from $10.00/M
Command A$10.00
Pricing data from OpenRouter API · Updated June 2026 · Prices may change