▸ playground usage
new: genuinely free AI models are live — try surp/free + see live budgets · token-gating prototype · vote on SRP

playground

surp/best-chat --interactive

Pick a combo, send a message. The first request will return 402 with payment requirements — you'll sign with your wallet, and the completion will stream back. (Real USDC on Base mainnet — sign with your wallet.)

live resolutions

surp/valuedeepseek-v4.1-flash$0.0012/1M-100%168 models

capable models (AA score within 20% of the best), then max Surplus discount vs AA list price — compares 168 model(s), routes to the cheapest. click any row to see the full list.

comboroutes toUSD / 1M tokvs $9/1M listpool
#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-coder-30b-a3b-instructcoding$0.003514
7qwen3-32bchat$0.003610
8openai-gpt-oss-120bchat$0.003759
9openai-gpt-oss-20bchat$0.003716
10nvidia-nemotron-3-nano-30b-a3bfast$0.003756
… and 158 more model(s) not shown
surp/frontierdeepseek-v4.1-flash$0.0012/1M-100%168 models

highest Artificial Analysis Intelligence Index, then Surplus discount — compares 168 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-coder-30b-a3b-instructcoding$0.003514
7qwen3-32bchat$0.003610
8openai-gpt-oss-120bchat$0.003759
9openai-gpt-oss-20bchat$0.003716
10nvidia-nemotron-3-nano-30b-a3bfast$0.003756
… and 158 more model(s) not shown
surp/fastdeepseek-v4.1-flash$0.0012/1M-100%168 models

highest AA median tokens/sec inside the 20% quality band — compares 168 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-coder-30b-a3b-instructcoding$0.003514
7qwen3-32bchat$0.003610
8openai-gpt-oss-120bchat$0.003759
9openai-gpt-oss-20bchat$0.003716
10nvidia-nemotron-3-nano-30b-a3bfast$0.003756
… and 158 more model(s) not shown
surp/visionpalmyra-vision-7b$0.0150/1M-100%5 models

multimodal / vision pool, then the value rule — compares 5 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1palmyra-vision-7b ◀ routed nowvision$0.01509
2qwen3-vl-235b-a22b-instructvision$0.021115
3qwen3-vl-30b-a3b-thinkingvision$0.209036
4qwen3-vl-235b-a22b-thinkingvision$0.400948
5glm-5v-turbovision$1.905826
surp/customdeepseek-v4.1-flash$0.0012/1M-100%168 models

your intelligence / speed / discount mix (surp_weights=cost:intel:speed) — compares 168 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-coder-30b-a3b-instructcoding$0.003514
7qwen3-32bchat$0.003610
8openai-gpt-oss-120bchat$0.003759
9openai-gpt-oss-20bchat$0.003716
10nvidia-nemotron-3-nano-30b-a3bfast$0.003756
… and 158 more model(s) not shown
surp/best-codingqwen3-coder-30b-a3b-instruct$0.0035/1M-100%7 models

cheapest coding-class model (coder / codex / qwen3-coder) — compares 7 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1qwen3-coder-30b-a3b-instruct ◀ routed nowcoding$0.003514
2kimi-k2.7-codecoding$0.005966
3qwen3-coder-nextcoding$0.009212
4qwen3-codercoding$0.013036
5qwen3-coder-turbocoding$0.462528
6gpt-5.3-codexcoding$0.783057
7gpt-5.2-codexcoding$3.937526
surp/best-reasoningkimi-k2-thinking$0.0310/1M-100%3 models

cheapest reasoning model (thinking / r1) — compares 3 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1kimi-k2-thinking ◀ routed nowreasoning$0.031020
2qwen3-235b-a22b-thinking-2507reasoning$0.096045
3deepseek-r1reasoning$2.24005
surp/best-fastgemma-3-4b-it$0.0015/1M-100%42 models

cheapest small/fast model (mini / nano / lite / small) — compares 42 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1gemma-3-4b-it ◀ routed nowfast$0.001514
2gemma-3-12b-itfast$0.002012
3gemma-4-e2bfast$0.00245
4qwen3-coder-30b-a3b-instructcoding$0.003514
5nvidia-nemotron-3-nano-30b-a3bfast$0.003756
6ministral-3-3b-instructfast$0.004010
7gemma-4-26b-a4b-itfast$0.005366
8nvidia-nemotron-nano-9b-v2fast$0.005810
9ministral-3-8b-instructfast$0.006010
10ministral-3-14b-instructfast$0.008010
… and 32 more model(s) not shown
surp/best-visionpalmyra-vision-7b$0.0150/1M-100%5 models

cheapest multimodal vision model (-vl / vision / 5v) — compares 5 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1palmyra-vision-7b ◀ routed nowvision$0.01509
2qwen3-vl-235b-a22b-instructvision$0.021115
3qwen3-vl-30b-a3b-thinkingvision$0.209036
4qwen3-vl-235b-a22b-thinkingvision$0.400948
5glm-5v-turbovision$1.905826
surp/best-chatdeepseek-v4.1-flash$0.0012/1M-100%153 models

cheapest general text LLM (no specialization) — compares 153 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-32bchat$0.003610
7openai-gpt-oss-120bchat$0.003759
8openai-gpt-oss-20bchat$0.003716
9nvidia-nemotron-3-nano-30b-a3bfast$0.003756
10ministral-3-3b-instructfast$0.004010
… and 143 more model(s) not shown
surp/best-coding-fastqwen3-coder-30b-a3b-instruct$0.0035/1M-100%1 models

cheapest coding model, biased to small/fast variants — compares 1 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1qwen3-coder-30b-a3b-instruct ◀ routed nowcoding$0.003514
surp/pro-codingqwen3-coder-30b-a3b-instruct$0.0035/1M-100%6 models

cheapest FRONTIER-tier coding model — compares 6 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1qwen3-coder-30b-a3b-instruct ◀ routed nowcoding$0.003514
2qwen3-coder-nextcoding$0.009212
3qwen3-codercoding$0.013036
4qwen3-coder-turbocoding$0.462528
5gpt-5.3-codexcoding$0.783057
6gpt-5.2-codexcoding$3.937526
surp/pro-reasoningdeepseek-r1$2.2400/1M-75%1 models

cheapest FRONTIER-tier reasoning model — compares 1 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-r1 ◀ routed nowreasoning$2.24005
surp/pro-visionpalmyra-vision-7b$0.0150/1M-100%5 models

cheapest FRONTIER-tier vision model — compares 5 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1palmyra-vision-7b ◀ routed nowvision$0.01509
2qwen3-vl-235b-a22b-instructvision$0.021115
3qwen3-vl-30b-a3b-thinkingvision$0.209036
4qwen3-vl-235b-a22b-thinkingvision$0.400948
5glm-5v-turbovision$1.905826
surp/pro-chatgrok-4.3$0.0375/1M-100%37 models

cheapest FRONTIER-tier chat model — compares 37 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1grok-4.3 ◀ routed nowchat$0.037554
2gpt-5.6-lunachat$0.048040
3gpt-5.6-luna-prochat$0.350027
4grok-4.20-betachat$0.375050
5gpt-5.5chat$0.385778
6gpt-5.6-terrachat$0.399762
7claude-sonnet-4.6chat$0.422562
8grok-4.1-fastchat$0.43755
9kimi-k3chat$0.476064
10grok-4.20-multi-agent-betachat$0.525041
… and 27 more model(s) not shown
surp/pro-fastqwen3-coder-30b-a3b-instruct$0.0035/1M-100%4 models

cheapest FRONTIER-tier fast model — compares 4 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1qwen3-coder-30b-a3b-instruct ◀ routed nowcoding$0.003514
2gemini-2.5-profast$0.666915
3gemini-3.1-pro-previewfast$0.734551
4gemini-3.1-profast$2.80004
surp/codingqwen3-coder-30b-a3b-instruct$0.0035/1M-100%7 models

alias of best-coding — compares 7 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1qwen3-coder-30b-a3b-instruct ◀ routed nowcoding$0.003514
2kimi-k2.7-codecoding$0.005966
3qwen3-coder-nextcoding$0.009212
4qwen3-codercoding$0.013036
5qwen3-coder-turbocoding$0.462528
6gpt-5.3-codexcoding$0.783057
7gpt-5.2-codexcoding$3.937526
surp/chatdeepseek-v4.1-flash$0.0012/1M-100%153 models

alias of best-chat — compares 153 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-32bchat$0.003610
7openai-gpt-oss-120bchat$0.003759
8openai-gpt-oss-20bchat$0.003716
9nvidia-nemotron-3-nano-30b-a3bfast$0.003756
10ministral-3-3b-instructfast$0.004010
… and 143 more model(s) not shown
surp/freedeepseek-v4.1-flash$0.0012/1M-100%168 models

treasury-sponsored free inference with live fallback and daily limits — compares 168 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-coder-30b-a3b-instructcoding$0.003514
7qwen3-32bchat$0.003610
8openai-gpt-oss-120bchat$0.003759
9openai-gpt-oss-20bchat$0.003716
10nvidia-nemotron-3-nano-30b-a3bfast$0.003756
… and 158 more model(s) not shown
surp/srup-freedeepseek-v4.1-flash$0.0012/1M-100%168 models

legacy alias of surp/free (treasury-sponsored) — compares 168 model(s), routes to the cheapest. click any row to see the full list.

#modelclassUSD / 1M toksellers
1deepseek-v4.1-flash ◀ routed nowchat$0.001239
2gemma-3-4b-itfast$0.001514
3gemma-3-12b-itfast$0.002012
4gemma-4-e2bfast$0.00245
5gemma-3-27b-itchat$0.002860
6qwen3-coder-30b-a3b-instructcoding$0.003514
7qwen3-32bchat$0.003610
8openai-gpt-oss-120bchat$0.003759
9openai-gpt-oss-20bchat$0.003716
10nvidia-nemotron-3-nano-30b-a3bfast$0.003756
… and 158 more model(s) not shown
surp/free-coding no models available
surp/free-fast no models available