Model Chooser
Which LLM Should I Use?
Filter models by price, context window, modality, and use case to find the right LLM, with a short recommendation for each top pick.
Updated FreeNo signup
DeepSeek V4 Flash
DeepSeek · $0.14 in / $0.28 out per 1M · 1M context. Good for reasoning / math, coding, long context, cheap / high-volume. It's the lowest total token price among 31 matching models.
| Model | In $/1M | Out $/1M | Context | Good for |
|---|---|---|---|---|
| DeepSeek V4 FlashDeepSeek | $0.14 | $0.28 | 1M | Reasoning / mathCodingLong contextCheap / high-volume |
| GPT-5 nanoOpenAI | $0.05 | $0.4 | 400K | Cheap / high-volume |
| GPT-4.1 nanoOpenAI | $0.1 | $0.4 | 1.0M | Cheap / high-volumeLong context |
| Gemini 2.5 Flash-LiteGoogle | $0.1 | $0.4 | 1.0M | Cheap / high-volume |
| GPT-4o miniOpenAI | $0.15 | $0.6 | 128K | Cheap / high-volumeVision / multimodal |
| Mistral Small 4Mistral | $0.15 | $0.6 | 128K | Cheap / high-volume |
| CodestralMistral | $0.3 | $0.9 | 256K | Coding |
| Gemini 3.1 Flash-LiteGoogle | $0.25 | $1.5 | 1.0M | Cheap / high-volume |
| GPT-4.1 miniOpenAI | $0.4 | $1.6 | 1.0M | Cheap / high-volumeLong context |
| Mistral Large 3Mistral | $0.5 | $1.5 | 128K | CodingAgents / tool use |
| Llama 3.3 70B Instruct TurboMeta | $1.04 | $1.04 | 131K | Cheap / high-volumeCoding |
| GPT-5 miniOpenAI | $0.25 | $2 | 400K | Cheap / high-volumeCoding |
| Gemini 2.5 FlashGoogle | $0.3 | $2.5 | 1.0M | Cheap / high-volumeLong contextVision / multimodal |
| Grok 4.3xAI | $1.25 | $2.5 | 1M | Long contextAgents / tool use |
| o4-miniOpenAI | $1.1 | $4.4 | 200K | Reasoning / mathCheap / high-volume |
| Claude Haiku 4.5Anthropic | $1 | $5 | 200K | Cheap / high-volume |
| Llama 3.1 405B Instruct TurboMeta | $3.5 | $3.5 | 131K | Reasoning / mathCoding |
| GPT-4.1OpenAI | $2 | $8 | 1.0M | CodingLong context |
| o3OpenAI | $2 | $8 | 200K | Reasoning / mathCoding |
| Gemini 3.5 FlashGoogle | $1.5 | $9 | 1.0M | Cheap / high-volumeLong contextVision / multimodal |
| GPT-5OpenAI | $1.25 | $10 | 400K | CodingAgents / tool useReasoning / math |
| Gemini 2.5 ProGoogle | $1.25 | $10 | 1.0M | Reasoning / mathCodingLong contextVision / multimodal |
| GPT-4oOpenAI | $2.5 | $10 | 128K | Vision / multimodalWriting |
| Gemini 3.1 ProGoogle | $2 | $12 | 1.0M | Reasoning / mathCodingAgents / tool useLong contextVision / multimodalWriting |
| GPT-5.4OpenAI | $2.5 | $15 | 400K | CodingAgents / tool useReasoning / mathVision / multimodal |
| Claude Sonnet 5Anthropic | $3 | $15 | 1M | CodingAgents / tool useWritingLong contextCheap / high-volume |
| Claude Sonnet 4.6Anthropic | $3 | $15 | 1M | CodingAgents / tool useLong context |
| Claude Opus 4.8Anthropic | $5 | $25 | 1M | CodingAgents / tool useReasoning / mathWritingLong contextVision / multimodal |
| Claude Opus 4.7Anthropic | $5 | $25 | 1M | CodingAgents / tool useLong context |
| GPT-5.5OpenAI | $5 | $30 | 400K | CodingAgents / tool useReasoning / mathWritingVision / multimodal |
| Claude Fable 5Anthropic | $10 | $50 | 1M | Reasoning / mathAgents / tool useCodingWriting |
TL;DR
There's no single "best" LLM — pick by task and constraints. Filter the 31 models above by use case, minimum context, price, and modality, and take the cheapest that clears your quality bar (often a mid-tier model, not the flagship). The full comparison table — price, context, and what each model is good for — is below.
Full model comparison
Updated Jul 5, 2026| Model | Provider | In $/1M | Out $/1M | Context | Good for |
|---|---|---|---|---|---|
| GPT-5 nano | OpenAI | $0.05 | $0.4 | 400K | Cheap / high-volume |
| GPT-4.1 nano | OpenAI | $0.1 | $0.4 | 1.0M | Cheap / high-volume, Long context |
| GPT-4o mini | OpenAI | $0.15 | $0.6 | 128K | Cheap / high-volume, Vision / multimodal |
| GPT-5 mini | OpenAI | $0.25 | $2 | 400K | Cheap / high-volume, Coding |
| GPT-4.1 mini | OpenAI | $0.4 | $1.6 | 1.0M | Cheap / high-volume, Long context |
| o4-mini | OpenAI | $1.1 | $4.4 | 200K | Reasoning / math, Cheap / high-volume |
| GPT-5 | OpenAI | $1.25 | $10 | 400K | Coding, Agents / tool use, Reasoning / math |
| GPT-4.1 | OpenAI | $2 | $8 | 1.0M | Coding, Long context |
| o3 | OpenAI | $2 | $8 | 200K | Reasoning / math, Coding |
| GPT-5.4 | OpenAI | $2.5 | $15 | 400K | Coding, Agents / tool use, Reasoning / math, Vision / multimodal |
| GPT-4o | OpenAI | $2.5 | $10 | 128K | Vision / multimodal, Writing |
| GPT-5.5 | OpenAI | $5 | $30 | 400K | Coding, Agents / tool use, Reasoning / math, Writing, Vision / multimodal |
| Claude Haiku 4.5 | Anthropic | $1 | $5 | 200K | Cheap / high-volume |
| Claude Sonnet 5 | Anthropic | $3 | $15 | 1M | Coding, Agents / tool use, Writing, Long context, Cheap / high-volume |
| Claude Sonnet 4.6 | Anthropic | $3 | $15 | 1M | Coding, Agents / tool use, Long context |
| Claude Opus 4.8 | Anthropic | $5 | $25 | 1M | Coding, Agents / tool use, Reasoning / math, Writing, Long context, Vision / multimodal |
| Claude Opus 4.7 | Anthropic | $5 | $25 | 1M | Coding, Agents / tool use, Long context |
| Claude Fable 5 | Anthropic | $10 | $50 | 1M | Reasoning / math, Agents / tool use, Coding, Writing |
| Gemini 2.5 Flash-Lite | $0.1 | $0.4 | 1.0M | Cheap / high-volume | |
| Gemini 3.1 Flash-Lite | $0.25 | $1.5 | 1.0M | Cheap / high-volume | |
| Gemini 2.5 Flash | $0.3 | $2.5 | 1.0M | Cheap / high-volume, Long context, Vision / multimodal | |
| Gemini 2.5 Pro | $1.25 | $10 | 1.0M | Reasoning / math, Coding, Long context, Vision / multimodal | |
| Gemini 3.5 Flash | $1.5 | $9 | 1.0M | Cheap / high-volume, Long context, Vision / multimodal | |
| Gemini 3.1 Pro | $2 | $12 | 1.0M | Reasoning / math, Coding, Agents / tool use, Long context, Vision / multimodal, Writing | |
| DeepSeek V4 Flash | DeepSeek | $0.14 | $0.28 | 1M | Reasoning / math, Coding, Long context, Cheap / high-volume |
| Grok 4.3 | xAI | $1.25 | $2.5 | 1M | Long context, Agents / tool use |
| Mistral Small 4 | Mistral | $0.15 | $0.6 | 128K | Cheap / high-volume |
| Codestral | Mistral | $0.3 | $0.9 | 256K | Coding |
| Mistral Large 3 | Mistral | $0.5 | $1.5 | 128K | Coding, Agents / tool use |
| Llama 3.3 70B Instruct Turbo | Meta | $1.04 | $1.04 | 131K | Cheap / high-volume, Coding |
| Llama 3.1 405B Instruct Turbo | Meta | $3.5 | $3.5 | 131K | Reasoning / math, Coding |
How it works
The chooser filters the model data layer by your constraints (max input price, minimum context window, required modality, use case) and ranks matches by total token price (input + output per 1M). The top pick is the lowest-priced model that meets every filter.
Use-case tags are editorial and dated — they reflect general positioning as of 2026-07-05, not a benchmark score, and the fast-moving field means they go stale quickly. Prices and context windows are verified against each provider's pricing page. Treat this as a starting shortlist, then evaluate on your own task.
FAQ
- What's the best LLM for coding?
- As of the updated date, the strongest coding models are Claude Opus 4.8, GPT-5.4/5.5, Gemini 3.1 Pro, and Claude Sonnet 5 (near-Opus quality at lower cost). For cheap, high-volume code tasks, GPT-4.1-mini, Gemini 2.5 Flash, and DeepSeek V4 are strong value. Filter by the 'Coding' use case above to compare price and context.
- What's the best LLM for building agents?
- Agentic workloads reward strong tool-use and long-horizon reasoning: Claude Opus 4.8, GPT-5.4, Gemini 3.1 Pro, and Claude Sonnet 5 lead here. Because agents accumulate context and make many calls, weigh price heavily — a cheaper model that's 'good enough' per step can win on total cost. Use the 'Agents / tool use' filter.
- Which model has the best price for high-volume use?
- For cheap, high-throughput work the standouts are GPT-5-nano, GPT-4.1-nano, Gemini 2.5 Flash-Lite, Claude Haiku 4.5, and DeepSeek V4 Flash — all well under $1 per 1M input tokens. Filter by 'Cheap / high-volume' and sort by price to see the current cheapest options.
- How should I choose between models?
- Start from the task (coding, agents, reasoning, long context, cheap high-volume, vision) and your hard constraints (max price, minimum context window, required modality). Filter to the models that qualify, then pick the cheapest that clears your quality bar — often a mid-tier model, not the flagship. Prices and benchmarks move, so re-check periodically.