talus protocol
Models and prices.
216 models. You pay OpenRouter's price x 1.20 for the tokens actually used, never more than the cap you set aside for the call. A chat@1 call costs at least 0.005 USDC.
How a call is priced
Each call costs the real model price plus 20%. Answers are limited to 1000 output tokens (about 750 words),2000 for the models with a measured throughput floor (deepseek/deepseek-v4.1-flash, deepseek/deepseek-v4.1-flash-1017), or less when the model's own limit is lower. A call also costs a small Talus Protocol network fee in SUI, paid from your own wallet; that is separate from the prices on this page.
You set aside a cap for each call. It must cover the worst case, so the whole answer fits. Whatever the answer does not cost comes back. The cap is at most 1 USDC. The last column shows the smallest cap for a short question (about 200 characters) and a full-length answer. The worst case is computed like this:
max_cost = ceil((prompt_bound x max(input, cache_write, 2 x input) + request_fee + T x max(output, reasoning)) x 1.2 x 1e6), prompt_bound = request_bytes + 64 + 32 + 16 x messages (one token per byte, the proxy's own upper bound), T = min(time ceiling, model output limit), time ceiling = output_cap_tokens (measured_output_cap_tokens for a model with a measured throughput floor); a cap below this can cut the answer short or be refused, a cap above 1 USDC is not accepted.
Prices change upstream; this list was captured on 2026-10-19.
Models
| Model | You pay per 1M tokens in | You pay per 1M tokens out | Max answer (tokens) | Cap for a short question (micro-USDC) |
|---|---|---|---|---|
| ~anthropic/claude-haiku-latest | $0.12 (up to $0.6) | $0.6 (up to $3) | 1000 | 742 |
| ~deepseek/deepseek-pro-latest | $0.27 | $9.6 | 1000 | 9,919 |
| ~openai/gpt-mini-latest | $0.9 | $5.4 | 1000 | 6,450 |
| ~unbiased/pareto-latest | $0.96 | $3.84 | 1000 | 4,960 |
| aion-labs/aion-rp-llama-3.1-8b | $0.96 | $1.92 | 1000 | 3,053 |
| amazon/nova-2-lite-v1 | $0.36 | $3 | 1000 | 3,419 |
| amazon/nova-lite-v1 | $0.072 | $0.288 | 1000 | 372 |
| amazon/nova-micro-v1 | $0.042 | $0.168 | 1000 | 217 |
| amazon/nova-premier-v1 | $3 | $15 | 1000 | 18,492 |
| amazon/nova-pro-v1 | $0.96 | $3.84 | 1000 | 4,950 |
| anthracite-org/magnum-v4-72b | $3 | $6 | 1000 | 9,529 |
| anthropic/claude-haiku-4.5 | $1.2 | $6 | 1000 | 7,407 |
| anthropic/claude-haiku-4.5:batch | $0.6 | $3 | 1000 | 3,711 |
| anthropic/claude-haiku-5.5 | $0.12 (up to $0.6) | $0.6 (up to $3) | 1000 | 741 |
| anthropic/claude-haiku-5.5:batch | $0.06 (up to $0.3) | $0.3 (up to $1.5) | 1000 | 372 |
| anthropic/claude-opus-4.1 | $18 | $90 | 1000 | 111,060 |
| anthropic/claude-opus-4.1:batch | $9 | $45 | 1000 | 55,638 |
| anthropic/claude-opus-4.5 | $6 | $30 | 1000 | 37,021 |
| anthropic/claude-opus-4.5:batch | $3 | $15 | 1000 | 18,546 |
| anthropic/claude-opus-4.6 | $6 | $30 | 1000 | 37,021 |
| anthropic/claude-opus-4.6:batch | $3 | $15 | 1000 | 18,546 |
| anthropic/claude-opus-4.7 | $6 | $30 | 1000 | 37,021 |
| anthropic/claude-opus-4.7:batch | $3 | $15 | 1000 | 18,546 |
| anthropic/claude-opus-4.8 | $6 | $30 | 1000 | 37,021 |
| anthropic/claude-opus-4.8:batch | $3 | $15 | 1000 | 18,546 |
| anthropic/claude-sonnet-4.5 | $3.6 (up to $7.2) | $18 (up to $27) | 1000 | 22,227 |
| anthropic/claude-sonnet-4.5:batch | $1.8 (up to $3.6) | $9 (up to $13.5) | 1000 | 11,135 |
| anthropic/claude-sonnet-4.6 | $3.6 | $18 | 1000 | 22,227 |
| anthropic/claude-sonnet-4.6:batch | $1.8 | $9 | 1000 | 11,135 |
| bytedance-seed/seed-1.6 | $0.3 (up to $0.6) | $2.4 (up to $4.8) | 1000 | 2,750 |
| bytedance-seed/seed-1.6-flash | $0.09 (up to $0.12) | $0.36 (up to $0.96) | 1000 | 467 |
| bytedance-seed/seed-2-1-turbo | $0.6 | $3 | 1000 | 3,707 |
| bytedance-seed/seed-2.0-code | $0.6 (up to $1.2) | $3.6 (up to $7.2) | 1000 | 4,306 |
| bytedance-seed/seed-2.0-lite | $0.3 (up to $0.6) | $2.4 (up to $4.8) | 1000 | 2,753 |
| bytedance-seed/seed-2.0-mini | $0.12 (up to $0.24) | $0.48 (up to $0.96) | 1000 | 622 |
| bytedance/ui-tars-1.5-7b | $0.12 | $0.24 | 1000 | 381 |
| cognitivecomputations/dolphin-mistral-24b-venice-edition | $0.24 | $1.08 | 1000 | 1,376 |
| cohere/command-a | $3 | $12 | 1000 | 15,457 |
| cohere/command-a-plus | $0.36 | $1.8 | 1000 | 2,219 |
| cohere/command-r-08-2024 | $0.18 | $0.72 | 1000 | 931 |
| cohere/command-r-plus-08-2024 | $3 | $12 | 1000 | 15,535 |
| cohere/command-r7b-12-2024 | $0.045 | $0.18 | 1000 | 233 |
| deepseek/deepseek-chat | $0.384 | $1.068 | 1000 | 1,515 |
| deepseek/deepseek-chat-v3-0324 | $0.348 | $1.368 | 1000 | 1,779 |
| deepseek/deepseek-chat-v3.1 | $0.3 | $1.14 | 1000 | 1,493 |
| deepseek/deepseek-v3.1-terminus | $0.324 | $1.2 | 1000 | 1,583 |
| deepseek/deepseek-v3.2 | $0.336 | $0.504 | 1000 | 896 |
| deepseek/deepseek-v3.2-exp | $0.324 | $0.492 | 1000 | 872 |
| deepseek/deepseek-v4-flash | $0.1092 | $0.2184 | 1000 | 347 |
| deepseek/deepseek-v4-pro | $1.14631 | $2.29262 | 1000 | 3,632 |
| deepseek/deepseek-v4-pro-0813 | $0.792 | $2.376 | 1000 | 3,309 |
| deepseek/deepseek-v4.1-flash | $0.05376 | $0.096 | 2000 | 263 |
| deepseek/deepseek-v4.1-flash-1017 | $0.36 | $1.44 | 2000 | 3,357 |
| elorian/glimpse | $0.72 | $3.6 | 1000 | 4,429 |
| google/gemini-2.5-flash | $0.36 | $3 | 1000 | 3,420 |
| google/gemini-2.5-flash-lite | $0.12 | $0.48 | 1000 | 622 |
| google/gemini-2.5-flash-lite:batch | $0.06 | $0.24 | 1000 | 312 |
| google/gemini-2.5-flash:batch | $0.18 | $1.5 | 1000 | 1,713 |
| google/gemini-3-flash-preview | $0.6 | $3.6 | 1000 | 4,307 |
| google/gemini-3-flash-preview:batch | $0.3 | $1.8 | 1000 | 2,157 |
| google/gemma-2-27b-it | $0.78 | $0.78 | 1000 | 1,687 |
| google/gemma-3-12b-it | $0.06 | $0.18 | 1000 | 250 |
| google/gemma-3-27b-it | $0.096 | $0.42 | 1000 | 532 |
| google/gemma-3-4b-it | $0.06 | $0.12 | 1000 | 190 |
| google/gemma-4-26b-a4b-it | $0.108 | $0.36 | 1000 | 487 |
| google/gemma-4-31b-it | $0.24 | $0.48 | 1000 | 759 |
| gryphe/mythomax-l2-13b | $0.096 | $0.132 | 1000 | 244 |
| ibm-granite/granite-4.0-h-micro | $0.0204 | $0.1344 | 1000 | 159 |
| inference-net/schematron-v2-small | $0.06 | $0.276 | 1000 | 348 |
| inference-net/schematron-v2-turbo | $0.036 | $0.18 | 1000 | 223 |
| mancer/weaver | $0.48 | $0.9 | 1000 | 1,451 |
| meta-llama/llama-3.1-70b-instruct | $0.48 | $0.48 | 1000 | 1,050 |
| meta-llama/llama-3.1-8b-instruct | $0.06 | $0.096 | 1000 | 168 |
| meta-llama/llama-3.2-1b-instruct | $0.0324 | $0.2412 | 1000 | 280 |
| meta-llama/llama-3.2-3b-instruct | $0.06108 | $0.402 | 1000 | 475 |
| meta-llama/llama-3.3-70b-instruct | $0.12 | $0.384 | 1000 | 527 |
| meta-llama/llama-4-maverick | $0.225 | $0.783 | 1000 | 1,048 |
| meta-llama/llama-4-scout | $0.12 | $0.36 | 1000 | 501 |
| meta-llama/llama-guard-4-12b | $0.216 | $0.216 | 1000 | 471 |
| microsoft/phi-4 | $0.084 | $0.168 | 1000 | 265 |
| microsoft/wizardlm-2-8x22b | $0.744 | $0.744 | 1000 | 1,616 |
| minimax/minimax-01 | $0.24 | $1.32 | 1000 | 1,598 |
| minimax/minimax-m1 | $0.66 | $2.64 | 1000 | 3,403 |
| minimax/minimax-m2-her | $0.36 | $1.44 | 1000 | 1,860 |
| minimax/minimax-m3 | $0.36 | $1.44 | 1000 | 1,857 |
| mistralai/codestral-2508 | $0.36 | $1.08 | 1000 | 1,501 |
| mistralai/codestral-2508:batch | $0.18 | $0.54 | 1000 | 753 |
| mistralai/devstral-2512 | $0.48 | $2.4 | 1000 | 2,960 |
| mistralai/ministral-14b-2512 | $0.24 | $0.24 | 1000 | 523 |
| mistralai/ministral-3b-2512 | $0.12 | $0.12 | 1000 | 261 |
| mistralai/ministral-8b-2512 | $0.18 | $0.18 | 1000 | 392 |
| mistralai/ministral-8b-2512:batch | $0.09 | $0.09 | 1000 | 197 |
| mistralai/mistral-large | $2.4 | $7.2 | 1000 | 9,999 |
| mistralai/mistral-large-2407 | $2.4 | $7.2 | 1000 | 10,023 |
| mistralai/mistral-large-2512 | $0.6 | $1.8 | 1000 | 2,506 |
| mistralai/mistral-large-2512:batch | $0.3 | $0.9 | 1000 | 1,257 |
| mistralai/mistral-medium-3 | $0.48 | $2.4 | 1000 | 2,963 |
| mistralai/mistral-medium-3-5 | $1.8 | $9 | 1000 | 11,117 |
| mistralai/mistral-medium-3-5:batch | $0.9 | $4.5 | 1000 | 5,570 |
| mistralai/mistral-medium-3.1 | $0.48 | $2.4 | 1000 | 2,965 |
| mistralai/mistral-medium-3.1:batch | $0.24 | $1.2 | 1000 | 1,486 |
| mistralai/mistral-nemo | $0.0348 | $0.036 | 1000 | 77 |
| mistralai/mistral-saba | $0.24 | $0.72 | 1000 | 1,000 |
| mistralai/mistral-small-24b-instruct-2501 | $0.06 | $0.096 | 1000 | 169 |
| mistralai/mistral-small-2603 | $0.18 | $0.72 | 1000 | 932 |
| mistralai/mistral-small-2603:batch | $0.09 | $0.36 | 1000 | 467 |
| mistralai/mistral-small-3.1-24b-instruct | $0.4212 | $0.666 | 1000 | 1,172 |
| mistralai/mistral-small-3.2-24b-instruct | $0.1125 | $0.3 | 1000 | 435 |
| mistralai/mixtral-8x22b-instruct | $2.4 | $7.2 | 1000 | 10,042 |
| mistralai/voxtral-small-24b-2507 | $0.12 | $0.36 | 1000 | 503 |
| moondream/moondream3-preview | $0.36 | $3 | 1000 | 3,424 |
| moondream/moondream3.1-9b-a2b | $0.36 | $3 | 1000 | 3,425 |
| moonshotai/kimi-k2 | $0.684 | $2.76 | 1000 | 3,551 |
| moonshotai/kimi-k2-0905 | $0.72 | $3 | 1000 | 3,840 |
| morph/morph-v3-fast | $0.96 | $1.44 | 1000 | 2,552 |
| morph/morph-v3-large | $1.08 | $2.28 | 1000 | 3,533 |
| nex-agi/nex-n2.5-mini | $0.03 | $0.12 | 1000 | 155 |
| nex-agi/nex-n2.5-pro | $0.09 | $0.3 | 1000 | 405 |
| nousresearch/hermes-3-llama-3.1-405b | $1.2 | $1.2 | 1000 | 2,631 |
| nousresearch/hermes-3-llama-3.1-70b | $0.84 | $0.84 | 1000 | 1,840 |
| nousresearch/hermes-4-405b | $1.2 | $3.6 | 1000 | 5,007 |
| nvidia/nemotron-3-nano-30b-a3b | $0.072 | $0.288 | 1000 | 373 |
| nvidia/nemotron-3.5-lightning | $0.084 | $0.24 | 1000 | 339 |
| nvidia/nemotron-3.5-super-vl | $0.12 | $0.48 | 1000 | 622 |
| openai/gpt-3.5-turbo | $0.6 | $1.8 | 1000 | 2,497 |
| openai/gpt-3.5-turbo-0613 | $1.2 | $2.4 | 1000 | 3,804 |
| openai/gpt-3.5-turbo-16k | $3.6 | $4.8 | 1000 | 9,005 |
| openai/gpt-3.5-turbo-instruct | $1.8 | $2.4 | 1000 | 4,521 |
| openai/gpt-3.5-turbo:batch | $0.3 | $0.9 | 1000 | 1,252 |
| openai/gpt-4 | $36 | $72 | 1000 | 113,185 |
| openai/gpt-4-turbo | $12 | $36 | 1000 | 49,872 |
| openai/gpt-4-turbo:batch | $6 | $18 | 1000 | 25,009 |
| openai/gpt-4.1 | $2.4 | $9.6 | 1000 | 12,356 |
| openai/gpt-4.1-mini | $0.48 | $1.92 | 1000 | 2,476 |
| openai/gpt-4.1-mini:batch | $0.24 | $0.96 | 1000 | 1,241 |
| openai/gpt-4.1-nano | $0.12 | $0.48 | 1000 | 619 |
| openai/gpt-4.1-nano:batch | $0.06 | $0.24 | 1000 | 311 |
| openai/gpt-4.1:batch | $1.2 | $4.8 | 1000 | 6,192 |
| openai/gpt-4o | $3 | $12 | 1000 | 15,439 |
| openai/gpt-4o-2024-05-13 | $6 | $18 | 1000 | 25,009 |
| openai/gpt-4o-2024-08-06 | $3 | $12 | 1000 | 15,505 |
| openai/gpt-4o-2024-11-20 | $3 | $12 | 1000 | 15,505 |
| openai/gpt-4o-mini | $0.18 | $0.72 | 1000 | 929 |
| openai/gpt-4o-mini-2024-07-18 | $0.18 | $0.72 | 1000 | 933 |
| openai/gpt-5.1 | $1.5 | $12 | 1000 | 13,722 |
| openai/gpt-5.1:batch | $0.75 | $6 | 1000 | 6,870 |
| openai/gpt-5.2 | $2.1 | $16.8 | 1000 | 19,211 |
| openai/gpt-5.2-chat | $2.1 | $16.8 | 1000 | 19,232 |
| openai/gpt-5.2:batch | $1.05 | $8.4 | 1000 | 9,618 |
| openai/gpt-5.3-codex | $2.1 | $16.8 | 1000 | 19,236 |
| openai/gpt-5.4 | $3 (up to $6) | $18 (up to $27) | 1000 | 21,444 |
| openai/gpt-5.4-mini | $0.9 | $5.4 | 1000 | 6,443 |
| openai/gpt-5.4-mini:batch | $0.45 | $2.7 | 1000 | 3,227 |
| openai/gpt-5.4-nano | $0.24 | $1.5 | 1000 | 1,778 |
| openai/gpt-5.4-nano:batch | $0.12 | $0.75 | 1000 | 891 |
| openai/gpt-5.4:batch | $1.5 (up to $3) | $9 (up to $13.5) | 1000 | 10,741 |
| openai/gpt-chat-latest | $6 | $36 | 1000 | 42,984 |
| perceptron/perceptron-mk1 | $0.18 | $1.8 | 1000 | 2,011 |
| perceptron/perceptron-mk1.5 | $0.18 | $1.8 | 1000 | 2,012 |
| qwen/qwen-2.5-72b-instruct | $0.432 | $0.48 | 1000 | 987 |
| qwen/qwen-2.5-7b-instruct | $0.12 | $0.24 | 1000 | 381 |
| qwen/qwen-2.5-coder-32b-instruct | $0.792 | $1.2 | 1000 | 2,138 |
| qwen/qwen-plus | $0.312 (up to $0.936) | $0.936 (up to $2.808) | 1000 | 1,295 |
| qwen/qwen2.5-vl-72b-instruct | $0.96 | $1.2 | 1000 | 2,329 |
| qwen/qwen3-14b | $0.144 | $0.288 | 1000 | 454 |
| qwen/qwen3-235b-a22b-2507 | $0.108 | $0.66 | 1000 | 787 |
| qwen/qwen3-30b-a3b-instruct-2507 | $0.108 | $0.36 | 1000 | 488 |
| qwen/qwen3-32b | $0.096 | $0.336 | 1000 | 447 |
| qwen/qwen3-coder | $0.36 | $1.2 | 1000 | 1,615 |
| qwen/qwen3-coder-30b-a3b-instruct | $0.084 | $0.336 | 1000 | 436 |
| qwen/qwen3-coder-flash | $0.234 (up to $0.624) | $1.17 (up to $3.12) | 1000 | 1,443 |
| qwen/qwen3-coder-next | $0.144 | $0.96 | 1000 | 1,128 |
| qwen/qwen3-next-80b-a3b-instruct | $0.108 | $1.32 | 1000 | 1,448 |
| qwen/qwen3-vl-235b-a22b-instruct | $0.252 | $2.28 | 1000 | 2,579 |
| qwen/qwen3-vl-30b-a3b-instruct | $0.18 | $0.72 | 1000 | 933 |
| qwen/qwen3-vl-8b-instruct | $0.3 | $0.9 | 1000 | 1,251 |
| qwen/qwen3.5-122b-a10b | $0.312 | $2.496 | 1000 | 2,860 |
| qwen/qwen3.5-27b | $0.324 | $2.592 | 1000 | 2,966 |
| qwen/qwen3.5-35b-a3b | $0.18 | $1.2 | 1000 | 1,409 |
| qwen/qwen3.5-397b-a17b | $0.6 | $4.32 | 1000 | 5,019 |
| qwen/qwen3.5-9b | $0.12 | $0.18 | 1000 | 318 |
| qwen/qwen3.5-flash-02-23 | $0.078 | $0.312 | 1000 | 404 |
| qwen/qwen3.5-plus-02-15 | $0.312 (up to $0.39) | $1.872 (up to $2.34) | 1000 | 2,236 |
| qwen/qwen3.5-plus-20260420 | $0.36 (up to $0.45) | $2.16 (up to $2.7) | 1000 | 2,582 |
| qwen/qwen3.6-flash | $0.225 (up to $0.9) | $1.35 (up to $3.6) | 1000 | 1,611 |
| qwen/qwen3.6-plus | $0.39 (up to $1.56) | $2.34 (up to $4.68) | 1000 | 2,791 |
| rekaai/reka-edge | $0.12 | $0.12 | 1000 | 259 |
| relace/relace-apply-3 | $1.02 | $1.5 | 1000 | 2,686 |
| relace/relace-search | $1.2 | $3.6 | 1000 | 4,993 |
| sao10k/l3-lunaris-8b | $0.048 | $0.06 | 1000 | 116 |
| sao10k/l3.1-euryale-70b | $1.02 | $1.02 | 1000 | 2,210 |
| tencent/hunyuan-a13b-instruct | $0.168 | $0.684 | 1000 | 882 |
| tencent/hy-mt2-1.8b | $0.0528 | $0.2124 | 1000 | 274 |
| tencent/hy-mt2-30b-a3b | $0.0888 | $0.354 | 1000 | 458 |
| tencent/hy-mt2-7b | $0.0888 | $0.354 | 1000 | 457 |
| thedrummer/cydonia-24b-v4.1 | $0.36 | $0.6 | 1000 | 1,023 |
| thedrummer/skyfall-36b-v2 | $0.66 | $0.96 | 1000 | 1,733 |
| thedrummer/unslopnemo-12b | $0.48 | $0.48 | 1000 | 1,042 |
| unbiased/pareto-26.10 | $0.96 | $3.84 | 1000 | 4,956 |
| unbiased/pareto-26.9 | $3 | $9 | 1000 | 12,481 |
| undi95/remm-slerp-l2-13b | $0.42 | $0.78 | 1000 | 1,271 |
| upstage/solar-mini4 | $0.06 | $0.24 | 1000 | 310 |
| upstage/solar-pro-3 | $0.18 | $0.72 | 1000 | 929 |
| upstage/solar-pro4 | $0.108 | $0.432 | 1000 | 557 |
| writer/palmyra-x5 | $0.72 | $7.2 | 1000 | 8,031 |
| x-ai/grok-4.20 | $1.5 (up to $3) | $3 (up to $6) | 1000 | 4,722 |
| xiaomi/mimo-v2.5 | $0.168 | $0.336 | 1000 | 530 |
| xiaomi/mimo-v2.5-pro | $0.522 | $1.044 | 1000 | 1,650 |
| xiaomi/mimo-v2.6-flash | $0.168 | $0.336 | 1000 | 532 |
| xiaomi/mimo-v2.6-pro | $0.522 | $1.044 | 1000 | 1,650 |
| xiaomi/mimo-v2.6-pro-ultraspeed | $5.22 | $10.44 | 1000 | 16,611 |
| z-ai/glm-4.5 | $0.72 | $2.64 | 1000 | 3,464 |
| z-ai/glm-4.5-air | $0.156 | $1.02 | 1000 | 1,200 |
| z-ai/glm-4.5v | $0.72 | $2.16 | 1000 | 2,986 |
| z-ai/glm-4.6 | $0.66 | $2.64 | 1000 | 3,396 |
| z-ai/glm-4.6v | $0.36 | $1.08 | 1000 | 1,493 |
"Up to" marks a model that charges a higher rate at peak hours or for very long prompts; the cap above already covers it. Models with a per-request fee include it in the cap.
Not served
- Free models (:free and the free router). They are rate-limited and produce no answer. (17 in this list)
- Models that may reason without being asked. Their thinking is not shown to stay inside the output cap, so they could cost more than any limit you set. (212 in this list)
- Models that search the web on their own. Each search adds a fee no cap can bound. (11 in this list)
- Models that return images or audio. Only text answers are served. (3 in this list)
- Models with no usable output price. (4 in this list)
- Routers and models whose price is not a fixed number (for example auto-routing), or that charge a fee the cap cannot account for. (7 in this list)
A model that is refused before it runs costs nothing from your USDC, but the network fee is still spent, so check the name here first.