Llama 3.3 70B Instruct against 49 models
Llama 3.3 70B Instruct costs $0.328 per 1,000 chat replies at list price. Here is every model teams spend the most on or move traffic to, sorted from cheapest, with a link to each side-by-side page.
| Model | Per 1,000 chat replies | To scale | Against Llama 3.3 70B Instruct | Confidence signal |
|---|---|---|---|---|
| gpt-oss-20b | $0.072 | 78% cheaper | yes | |
| Qwen3.7 Flash | $0.112 | 66% cheaper | yes | |
| Nova Micro 1.0 | $0.126 | 62% cheaper | no | |
| Command R7B (12-2024) | $0.135 | 59% cheaper | no | |
| DeepSeek V4 Flash 0731 | $0.17 | 48% cheaper | yes | |
| Phi 4 | $0.196 | 40% cheaper | no | |
| Nova Lite 1.0 | $0.216 | 34% cheaper | no | |
| Nemotron 3.5 Lightning | $0.24 | 27% cheaper | yes | |
| DeepSeek V4.1 Flash | $0.29 | 12% cheaper | yes | |
| Gemma 4 31B | $0.316 | same | yes | |
| Ministral 3 8B 2512 | $0.36 | 1.1× the price | no | |
| MiMo-V2.6-Flash | $0.392 | 1.2× the price | no | |
| GPT-6 Luna | $0.40 | 1.2× the price | no | |
| Qwen3.8 Flash | $0.488 | 1.5× the price | yes | |
| GLM 5.3 Flash | $0.50 | 1.5× the price | yes | |
| gpt-oss-120b | $0.54 | 1.6× the price | yes | |
| Mistral Small 4 | $0.54 | 1.6× the price | no | |
| Llama 4 Maverick | $0.636 | 1.9× the price | yes | |
| GPT-5.6 Luna | $0.88 | 2.7× the price | no | |
| GPT-5.4 Nano | $0.90 | 2.7× the price | no | |
| MiniMax M3 | $1.08 | 3.3× the price | yes | |
| GPT-4.1 Mini | $1.44 | 4.4× the price | no | |
| Mistral Large 3 2512 | $1.60 | 4.9× the price | no | |
| Gemini 3.5 Flash Lite | $1.60 | 4.9× the price | no | |
| Qwen3.8 27B | $2.04 | 6.2× the price | yes | |
| Nemotron 3 Ultra | $2.16 | 6.6× the price | no | |
| Kimi K2.6 | $2.66 | 8.1× the price | yes | |
| Hy4 preview | $2.67 | 8.1× the price | no | |
| DeepSeek V4 Pro 0423 | $2.67 | 8.2× the price | yes | |
| Gemini 3.8 Flash | $3.00 | 9.1× the price | no | |
| GPT-5.4 Mini | $3.30 | 10× the price | no | |
| Claude Haiku 4.5 | $4.00 | 12× the price | no | |
| GLM 5.3 | $4.56 | 14× the price | yes | |
| Grok 4.7 | $5.12 | 16× the price | yes | |
| Qwen3.8 Max (0902) | $6.40 | 20× the price | yes | |
| Claude Sonnet 5 | $8.00 | 24× the price | no | |
| GPT-6 Sol | $8.00 | 24× the price | no | |
| GPT-5.6 Sol | $8.00 | 24× the price | no | |
| GPT-5.6 Terra | $8.80 | 27× the price | no | |
| Gemini 3.1 Pro Preview | $8.80 | 27× the price | no | |
| Command A | $9.00 | 27× the price | no | |
| Nova Premier 1.0 | $10.00 | 30× the price | no | |
| GPT-5.4 | $11.00 | 34× the price | no | |
| Kimi K3 | $12.00 | 37× the price | yes | |
| Claude Opus 5.5 | $16.00 | 49× the price | no | |
| Claude Opus 5 | $20.00 | 61× the price | no | |
| GPT-5.5 | $22.00 | 67× the price | no | |
| Claude Fable 5.1 | $40.00 | 122× the price | no | |
| GPT-6 Astra | $40.00 | 122× the price | no |
Find the cheapest model that answers your Llama 3.3 70B Instruct traffic as well
A pilot tests the candidates on your own requests.
Request a pilot