SOFT CAT.ai
FIND SOMETHING USEFUL
← tools

Model Comparison

Put a workload next to a price. See the assumptions, compare the costs and keep a copy of the numbers.

Model data. Verified-price snapshots from . Details

35 of 39 tracked models have verified token rates in the saved catalogue. Rates come from the OpenRouter public API. Models missing from that catalogue have unknown pricing and are excluded from cost estimates and chat selection.

Rates are USD per million text tokens. Caching, reasoning tokens, images, tool charges, provider routing and long-context tiers can change the bill. Check the provider before budgeting.

Context limits have their own dated review. The Context Budget Planner shows the catalogue and top-provider quotes separately, using the smaller known limit. Output caps are a separate constraint. Weight links identify repository revisions whose file lists were checked. Licence labels come from the model cards and are not an open-source certification. Check the actual terms and access conditions. Published weights do not make a hosted API or self-hosting free.

Inspect the dated data · Bot records

01 / SET A WORKLOAD

What would your calls cost?

Calculated here · no API call
Illustrative presets

Every call uses the same counts. Include the system prompt, tools and any repeated conversation history in the input. For a growing conversation, use the turn-by-turn calculator.

02 / COMPARE THE SNAPSHOTS

A price is a starting point.

These are recorded OpenRouter text-token rates, in USD. Try the models on your own examples before choosing one. Cost does not measure answer quality.

39 of 39 models shown. 35 have an estimate for this workload.

Scroll the table sideways on smaller screens. Select a column heading to sort.

Recorded model rates and estimated workload costs. All amounts are US dollars. Saved context is reference data.
Price snapshot
GPT-5 NanoOpenAI
openai/gpt-5-nano
400,000Reply cap: 128,000$0.05$0.4$0.25Rates checkedSource ↗
Llama 4 ScoutMeta
meta-llama/llama-4-scout
Weight files · gated access ↗llama4 · checked 2026-09-13
327,680Reply cap: 16,384$0.1$0.3$0.25Rates checkedSource ↗
Nemotron 3 SuperNvidia
nvidia/nemotron-3-super-120b-a12b
Published weight files ↗nvidia-nemotron-open-model-license · checked 2026-09-13
262,144Reply cap: 16,384$0.08$0.45$0.305Rates checkedSource ↗
Mistral Small 4Mistral
mistralai/mistral-small-2603
Published weight files ↗apache-2.0 · checked 2026-09-13
262,144Reply cap: 209,715$0.15$0.6$0.45Rates checkedSource ↗
Llama 4 MaverickMeta
meta-llama/llama-4-maverick
Weight files · gated access ↗llama4 · checked 2026-09-13
128,000Reply cap: 115,200$0.1875$0.6525$0.51375Rates checkedSource ↗
Qwen 2.5 72BAlibaba
qwen/qwen-2.5-72b-instruct
Published weight files ↗qwen · checked 2026-09-13
32,768Reply cap: 16,384$0.36$0.4$0.56Rates checkedSource ↗
Mistral Small 3.1Mistral
mistralai/mistral-small-3.1-24b-instruct
Published weight files ↗apache-2.0 · checked 2026-09-13
128,000Reply cap: 102,400$0.351$0.555$0.6285Rates checkedSource ↗
Qwen3.6 35B A3BAlibaba
qwen/qwen3.6-35b-a3b
Published weight files ↗apache-2.0 · checked 2026-09-13
262,144Reply cap: 235,929$0.15$1$0.65Rates checkedSource ↗
Trinity Large ThinkingArcee-Ai
arcee-ai/trinity-large-thinking
Published weight files ↗openmdw-1.1 · checked 2026-09-13
262,144Reply cap: 80,000$0.25$0.8$0.65Rates checkedSource ↗
DeepSeek V3DeepSeek
deepseek/deepseek-chat-v3-0324
Published weight files ↗mit · checked 2026-09-13
163,840Reply cap: 147,456$0.25$1$0.75Rates checkedSource ↗
MiniMax M2Minimax
minimax/minimax-m2
Published weight files ↗modified-mit · checked 2026-09-13
204,800Reply cap: 131,072$0.255$1.02$0.765Rates checkedSource ↗
MiniMax M2.7Minimax
minimax/minimax-m2.7
Published weight files ↗Custom licence, see model card · checked 2026-09-13
204,800Reply cap: 131,072$0.3$1.2$0.90Rates checkedSource ↗
GPT-5 MiniOpenAI
openai/gpt-5-mini
400,000Reply cap: 128,000$0.25$2$1.25Rates checkedSource ↗
Qwen3.6 27BAlibaba
qwen/qwen3.6-27b
Published weight files ↗apache-2.0 · checked 2026-09-13
262,144Reply cap: 65,536$0.3$2$1.30Rates checkedSource ↗
Gemini 2.5 FlashGoogle
google/gemini-2.5-flash
1,048,576Reply cap: 65,535$0.3$2.5$1.55Rates checkedSource ↗
GLM-5Z.ai
z-ai/glm-5
Published weight files ↗mit · checked 2026-09-13
198,000Reply cap: 128,000$0.6$1.92$1.56Rates checkedSource ↗
Kimi k2Moonshot AI
moonshotai/kimi-k2
Published weight files ↗modified-mit · checked 2026-09-13
131,072Reply cap: 100,352$0.57$2.3$1.72Rates checkedSource ↗
DeepSeek R1DeepSeek
deepseek/deepseek-r1
Published weight files ↗mit · checked 2026-09-13
64,000Reply cap: 16,000$0.7$2.5$1.95Rates checkedSource ↗
GLM 5.1Z.ai
z-ai/glm-5.1
Published weight files ↗mit · checked 2026-09-13
200,000Reply cap: 128,000$0.966$3.036$2.484Rates checkedSource ↗
Grok 4.3xAI
x-ai/grok-4.3
1,000,000Reply cap: 900,000$1.25$2.5$2.50Rates checkedSource ↗
Kimi K2.6Moonshot AI
moonshotai/kimi-k2.6
Published weight files ↗modified-mit · checked 2026-09-13
262,144Reply cap: 235,929$0.95$4$2.95Rates checkedSource ↗
o3-miniOpenAI
openai/o3-mini
200,000Reply cap: 100,000$1.1$4.4$3.30Rates checkedSource ↗
Claude Haiku 4.5Anthropic
anthropic/claude-haiku-4.5
200,000Reply cap: 64,000$1$5$3.50Rates checkedSource ↗
Gemini 3.5 FlashGoogle
google/gemini-3.5-flash
1,048,576Reply cap: 65,536$1.5$9$6.00Rates checkedSource ↗
GPT-4.1OpenAI
openai/gpt-4.1
1,047,576Reply cap: 32,768$2$8$6.00Rates checkedSource ↗
o3OpenAI
openai/o3
200,000Reply cap: 100,000$2$8$6.00Rates checkedSource ↗
Gemini 2.5 ProGoogle
google/gemini-2.5-pro-preview
1,048,576Reply cap: 65,536$1.25$10$6.25Rates checkedSource ↗
GPT-5OpenAI
openai/gpt-5
400,000Reply cap: 128,000$1.25$10$6.25Rates checkedSource ↗
GPT-4oOpenAI
openai/gpt-4o
128,000Reply cap: 16,384$2.5$10$7.50Rates checkedSource ↗
GPT-5.4OpenAI
openai/gpt-5.4
1,050,000Reply cap: 128,000$2.5$15$10.00Rates checkedSource ↗
Claude Sonnet 4Anthropic
anthropic/claude-sonnet-4
200,000Reply cap: 64,000$3$15$10.50Rates checkedSource ↗
Claude Opus 4.7Anthropic
anthropic/claude-opus-4.7
1,000,000Reply cap: 128,000$5$25$17.50Rates checkedSource ↗
GPT-5.5OpenAI
openai/gpt-5.5
1,050,000Reply cap: 128,000$5$30$20.00Rates checkedSource ↗
Claude Fable 5Anthropic
anthropic/claude-fable-5
1,000,000Reply cap: 128,000$10$50$35.00Rates checkedSource ↗
GPT-5.5 ProOpenAI
openai/gpt-5.5-pro
1,050,000Reply cap: 128,000$30$180$120.00Rates checkedSource ↗
Claude Opus 4Anthropic
anthropic/claude-opus-4
200,000Reply cap: 32,000Not listedNot listedPrice unknownnot listedSource ↗
Gemini 2.0 FlashGoogle
google/gemini-2.0-flash-001
UnknownReply cap: UnknownNot listedNot listedPrice unknownnot listedSource ↗
Laguna XS.2Poolside
poolside/laguna-xs.2
Published weight files ↗apache-2.0 · checked 2026-09-13
UnknownReply cap: UnknownNot listedNot listedPrice unknownnot listedSource ↗
Mistral Large 3Mistral
mistralai/mistral-large-2512
Published weight files ↗apache-2.0 · checked 2026-09-13
262,144Reply cap: 209,715Not listedNot listedPrice unknownnot listedSource ↗

Input plus output must fit the smaller recorded context limit, and a known output cap is checked separately. These are dated catalogue quotes, not a promise that every provider route accepts the request. Inspect both source values in the Context Budget Planner. Weight links point to recorded repository revisions with published files. Licences vary, and gated files may require approval from their host. A missing weight record does not establish that a model is closed.

Caching, reasoning tokens, images, tools, routing and long-context pricing can change the bill. Unknown rates stay unknown, including for zero calls. Open weights do not make hosted calls or hardware free.