inferenstack — the market read on AI inference · last observation 2026-09-20. See the boards

Price spread

The same model priced across every provider, on the same basis. The spread ratio (max ÷ min output $/M) is the number that says whether you're overpaying — and who the outlier is.

Published output $/M vendor-claimed · base tier, daily snapshot.

604 models
ModelProvidersCheapest $/MMedianSpreadCacheTiers
Mistral Nemomistralopen· 16K ctx8$0.030kilo$0.1351,333×25%–
Voxtral Small 24B 2507mistralopen· 33K ctx7$0.0023evroc$0.300223×29%–
GLM-5.3-Flashzhipuaiopen· 1M ctx63$0.025tokengo$0.500179×89%–
GPT-5.5openai· 1M ctx47$1.13unorouter$30.0067×91%yes
DeepSeek OCR 2deepseekopen· 8K ctx2$0.030novita-ai$0.74949×50%–
GPT OSS 120Bopenaiopen· 131K ctx80$0.100synthetic$0.60042×36%–
Nemotron 3 Ultra 550B A55Bnvidiaopen· 1M ctx13$0.100routing-run$2.4036×92%–
modelsdev/gemini31flashimagemodelsdev· 131K ctx8$2.00requesty$31.5030×13%–
x-ai/grok-4-fastx-ai· 2M ctx7$0.500zenmux$15.0030×100%–
Claude Opus 4.8anthropic· 1M ctx52$2.13unorouter$25.0028×87%yes
Nano Bananagoogle· 33K ctx11$1.25kilo$2.5024×64%–
Qwen3.6 35B-A3Balibabaopen· 262K ctx28$0.420empiriolabs$1.2624×32%–
deepseek-ai/deepseek-ocrdeepseek-aiopen· 8K ctx2$0.030novita-ai$0.36523×0%–
meta-llama/llama-3.2-1b-instructmeta-llamaopen· 131K ctx5$0.010inference$0.20120×20%–
GPT OSS 20Bopenaiopen· 131K ctx43$0.090kilo$0.20020×35%–
Grok 4.1 Fastxai· 2M ctx3$0.500ofox$0.50020×67%–
Nano Banana 2 Litegoogle· 66K ctx6$1.50edenai$1.5020×17%–
DeepSeek V4 Flash 0731deepseekopen· 1.0M ctx53$0.070merge-gateway$0.29019×91%–
DeepSeek V4 Flashdeepseekopen· 1.0M ctx80$0.070merge-gateway$0.28019×89%–
Llama-3.1-8B-Instructmetaopen· 131K ctx18$0.025inference$0.12618×22%–
accounts/fireworks/routers/glm-flash-latestaccountsopen· 1.0M ctx4$0.250openrouter$0.37518×100%–
GLM-4.7zhipuaiopen· 203K ctx36$0.800nano-gpt$2.2018×72%yes
meta-llama/llama-3.2-3b-instructmeta-llamaopen· 131K ctx9$0.020inference$0.05017×22%–
DeepSeek-R1-Distill-Qwen-32Bdeepseekopen· 131K ctx3$0.300novita-ai$0.86116×0%–
GPT-5.6 Lunaopenai· 1.1M ctx44$0.370bothub$1.2016×93%yes
qwen/qwen-2.5-72b-instructqwenopen· 131K ctx9$0.380abacus$0.40815×11%–
GPT-5.2openai· 400K ctx31$2.00qihang-ai$14.0014×87%–
qwen/qwen2.5-7b-instructqwenopen· 33K ctx8$0.050siliconflow$0.17214×0%–
Qwen3.5 9Balibabaopen· 256K ctx21$0.130empiriolabs$0.15013×24%–
Llama-3.3-70B-Instructmetaopen· 131K ctx38$0.230nano-gpt$0.72013×16%–
GPT-5 Miniopenai· 400K ctx34$0.290qihang-ai$2.0012×91%–
Llama-3.2-11B-Vision-Instructmetaopen· 1.0M ctx3$0.055inference$0.34512×0%–
GPT-5.2 Codexopenai· 400K ctx20$1.14qihang-ai$14.0012×90%–
Qwen2.5-VL 72B Instructalibabaopen· 128K ctx8$0.700nano-gpt$1.0012×38%–
GPT-5.6 Terraopenai· 1.1M ctx43$2.00xpersona$12.0012×95%yes
qwen/qwen3-8bqwenopen· 41K ctx9$0.060siliconflow$0.28712×22%–
Qwen3.8 27Balibabaopen· 262K ctx36$0.350llmgateway-providers$2.5211×75%–
Qwen3 32Balibabaopen· 41K ctx23$0.250ovhcloud$0.57011×13%–
Gemini 3.5 Flashgoogle· 1.0M ctx37$1.11unorouter$9.0011×92%–
Qwen3 235B-A22B Instruct 2507alibabaopen· 262K ctx26$0.300submodel$0.67510×23%–
MiniMax-M2.7minimaxopen· 205K ctx58$0.320llmgateway-providers$1.2010×81%–
Nemotron 3 Super 120B A12Bnvidiaopen· 262K ctx12$0.250nano-gpt$0.85010×50%–
Solar Pro 4upstage· 524K ctx5$0.120nano-gpt$0.36010×80%–
modelsdev/gemini3proimagemodelsdev· 66K ctx9$12.00abacus$12.0010×67%–
modelsdev/qwenvlocrmodelsdev· 34K ctx2$0.072alibaba-cn$0.39610×0%–
modelsdev/qwenmtplusmodelsdev· 16K ctx3$0.750novita-ai$0.7759.8×0%–
GLM-5.2zhipuaiopen· 1.0M ctx101$1.05crof$4.409.8×88%–
qwen/qwen3-coderqwenopen· 262K ctx5$0.500nano-gpt$1.209.8×40%–
Gemma 4 31B ITgoogleopen· 262K ctx36$0.250llmgateway-providers$0.4009.7×39%–
Qwen3-Coder 480B-A35B Instructalibaba· 262K ctx21$0.800submodel$1.559.4×19%yes
Qwen3-Next 80B-A3B (Thinking)alibabaopen· 256K ctx14$0.650nano-gpt$1.209.2×7%–
Qwen3-Coder 30B-A3B Instructalibabaopen· 262K ctx21$0.245cortecs$0.3009.2×14%yes
Gemini 3 Flash Previewgoogle· 1.0M ctx30$0.430qihang-ai$3.008.7×90%yes
DeepSeek V4.1 Flashdeepseekopen· 1.0M ctx49$0.280amd$1.208.6×96%–
Claude Haiku 4.5anthropic· 200K ctx13$0.710qihang-ai$5.008.5×62%–
Kimi K2.6moonshotaiopen· 262K ctx72$1.10routing-run$4.008.1×82%–
Mistral Large (latest)mistralopen· 128K ctx8$1.50mistral$6.008.0×50%–
MiniMax-M2.1minimaxopen· 197K ctx22$1.10llmgateway$1.207.6×64%–
modelsdev/deepseekr1distillllama70bmodelsdevopen· 33K ctx7$0.130helicone$0.8007.6×0%–
Ornith 1.5 35B A3Bdeepreinforceopen· 262K ctx3$0.400nano-gpt$0.4007.5×100%–
modelsdev/mistralnemoinstruct2407modelsdevopen· 128K ctx6$0.040io-net$0.1197.5×50%–
Kimi K2.7 Codemoonshotaiopen· 262K ctx72$1.10routing-run$4.007.3×92%–
modelsdev/qwen3omniflashmodelsdev· 66K ctx2$0.230alibaba-cn$0.9457.2×0%–
MiMo-V2.5xiaomiopen· 1.1M ctx22$0.280cline-pass$0.2807.1×91%yes
MiMo-V2-Omnixiaomi· 265K ctx3$0.280xiaomi$2.007.1×100%–
GLM-5V-Turbozhipuai· 203K ctx17$3.10aihubmix$4.007.1×94%yes
Qwen3.7 Flashalibaba· 1M ctx12$0.113aihubmix$0.1307.1×100%yes
Claude Sonnet 4.5anthropic· 200K ctx10$2.14qihang-ai$15.007.0×60%–
Claude Opus 4.5anthropic· 200K ctx11$3.57qihang-ai$25.007.0×64%–
Qwen3.5 Plusalibaba· 1M ctx15$0.688orcarouter$2.407.0×53%yes
DeepSeek V4 Prodeepseekopen· 1.0M ctx76$0.696routing-run$3.106.9×89%–
Seed 1.8bytedance-seed· 256K ctx6$0.290ofox$2.006.9×100%yes
Seed 1.6bytedance-seed· 256K ctx9$0.290ofox$2.006.9×78%yes
MiMo-V2-Proxiaomi· 1M ctx5$0.870xiaomi$3.006.9×100%yes
DeepSeek V4 Flash Vision Expdeepseekopen· 1.0M ctx18$0.200crof$0.6536.6×94%–
GLM-5.3zhipuaiopen· 1M ctx65$1.40crof$4.406.4×91%–
DeepSeek V3.2deepseekopen· 164K ctx49$0.295orcarouter$0.4256.3×63%–
Muse Glimmer 30Bmetaopen· 131K ctx12$0.800empiriolabs$1.506.3×92%–
DeepSeek V4 Pro 0813deepseekopen· 1M ctx41$0.800crof$3.966.2×93%–
Qwen3.6 27Balibabaopen· 262K ctx26$0.600pioneer$3.006.0×42%–
Gemini 2.5 Flash-Litegoogle· 1.0M ctx24$0.100llmtr$0.4006.0×88%–
Llama-3.2-3Bmetaopen· 131K ctx2$0.100pioneer$0.3506.0×50%–
Qwen3 Maxalibaba· 256K ctx23$1.43ofox$6.005.9×61%yes
modelsdev/qwen3235ba22bthinking2507modelsdev· 262K ctx13$0.600siliconflow-cn$2.305.8×15%–
qwen/qwen3-14bqwenopen· 41K ctx8$0.240vercel$0.2805.8×13%–
Qwen3 235B-A22Balibabaopen· 41K ctx12$0.500nano-gpt$0.9745.7×17%–
Qwen-VL Maxalibaba· 131K ctx5$0.574alibaba-cn$0.5805.6×60%–
o3openai· 200K ctx21$7.20poe$8.005.6×90%–
MiniMax-M2.5minimaxopen· 197K ctx48$0.880dinference$1.205.5×81%–
MiMo-V2.5-Proxiaomiopen· 1.0M ctx25$0.800crof$0.8705.4×88%yes
modelsdev/gemini25flash0520modelsdev· 1.0M ctx2$0.600nano-gpt$1.885.3×50%–
DeepSeek-R1deepseekopen· 164K ctx22$1.70nano-gpt$2.645.1×18%–
Mistral Medium 3.5mistralopen· 262K ctx16$2.00merge-gateway$7.505.1×63%–
Qwen2.5-Coder-32B-Instructalibabaopen· 33K ctx7$0.200huggingface$1.005.0×14%–
GPT-5.4openai· 922K ctx43$6.00xpersona$15.005.0×95%yes
minimax/minimax-m2.1-lightningminimax· 197K ctx4$0.480llmgateway-providers$1.445.0×50%–
mistralai/devstral-small-2505mistralaiopen· 33K ctx3$0.060nano-gpt$0.2205.0×67%–
inception/mercury-2.5-previewinception· 260K ctx6$0.150vercel$0.1505.0×100%–
modelsdev/llama38binstructmodelsdevopen· 131K ctx2$0.040novita-ai$0.1195.0×0%–
Nemotron 3 Nano Omni 30B A3B Reasoningnvidiaopen· 256K ctx3$0.380vultr$0.8004.8×33%–
Qwen3.5 397B-A17Balibabaopen· 262K ctx36$1.03orcarouter$3.604.8×39%yes
x-ai/grok-4.20x-ai· 2M ctx5$2.50openrouter$2.834.8×100%yes
Gemma 3 27B ITgoogleopen· 8K ctx17$0.160huggingface$0.3004.8×29%–
Seed 2.0 Litebytedance-seed· 262K ctx7$0.534volcengine$2.004.7×57%yes
Qwen Maxalibaba· 33K ctx9$1.38alibaba-cn$6.394.6×56%–
deepseek/deepseek-v3.2-thinkingdeepseek· 128K ctx3$0.420nano-gpt$0.4304.4×33%–
Qwen3.5 35B-A3Balibabaopen· 262K ctx20$0.459orcarouter$1.304.4×50%yes
Qwen3 Coder Flashalibaba· 1M ctx13$0.574alibaba-cn$1.504.4×77%yes
Whisper 3 Largeopenaiopen· 448 ctx2$0.0023evroc$0.00624.3×0%–
mistralai/mistral-small-24b-instruct-2501mistralai· 33K ctx3$0.080kilo$0.0804.3×0%–
Qwen3.5 122B-A10Balibabaopen· 262K ctx20$0.917merge-gateway$2.364.3×35%yes
Qwen3.7 Plusalibaba· 1M ctx28$1.13aihubmix$1.604.3×86%yes
Qwen Plusalibaba· 131K ctx13$0.287alibaba-cn$1.204.2×85%yes
QwQ Plusalibaba· 131K ctx3$0.574alibaba-cn$2.404.2×0%–
Gemini 2.5 Flashgoogle· 1.0M ctx35$0.600nano-gpt$2.504.2×89%yes
sao10k/l3.1-70b-euryale-v2.2sao10kopen· 20K ctx2$0.357nano-gpt$0.9184.1×50%–
GLM-4.6V-Flashzhipuaiopen· 131K ctx2$0.218zenmux$0.5594.1×50%yes
sao10k/l3-8b-stheno-v3.2sao10kopen· 16K ctx2$0.050novita-ai$0.1254.0×50%–
Kimi K2 Thinkingmoonshotaiopen· 262K ctx11$2.00vercel$2.504.0×82%–
GPT-5.6 Solopenai· 1.1M ctx43$10.00cloudflare-ai-gateway$20.004.0×95%yes
Gemini 3.6 Flashgoogle· 1.0M ctx27$1.88kilo$3.754.0×96%–
meta-llama/llama-prompt-guard-2-86mmeta-llamaopen· 512 ctx2$0.010helicone$0.0254.0×0%–
Qwen3.5 Flashalibaba· 1M ctx10$0.260openrouter$0.4004.0×60%yes
Qwen3 Coder Plusalibaba· 1M ctx14$2.29merge-gateway$5.003.9×79%yes
Qwen3.5 27Balibabaopen· 262K ctx18$0.688empiriolabs$2.133.8×33%yes
GLM-4.5-Airzhipuaiopen· 131K ctx19$0.291zenmux$0.8503.8×79%yes
modelsdev/nemotron3nanoomnimodelsdevopen· 66K ctx2$0.240requesty$0.5703.8×50%–
Kimi K3moonshotaiopen· 1.0M ctx83$8.00crof$15.003.8×92%–
Qwen3 30B A3Balibabaopen· 41K ctx12$0.290helicone$0.5003.7×17%–
modelsdev/kimik2modelsdev· 128K ctx8$2.20fastrouter$2.303.6×25%–
Qwen3.7 Maxalibaba· 1M ctx34$2.48merge-gateway$7.503.6×91%yes
modelsdev/phi4modelsdevopen· 128K ctx4$0.140kilo$0.3203.6×0%–
Gemini 3 Pro Previewgoogle· 1.0M ctx8$3.43qihang-ai$12.003.5×63%yes
inclusionai/ling-3.0-flashinclusionaiopen· 262K ctx7$0.063vercel$0.1803.5×100%–
Qwen3-Next 80B-A3B Instructalibabaopen· 262K ctx20$0.574merge-gateway$1.203.5×15%–
Qwen-Omni Turboalibaba· 33K ctx4$0.230alibaba-cn$0.5353.5×0%–
Nemotron 3 Nano 30B A3Bnvidiaopen· 262K ctx9$0.200edenai$0.2003.4×78%–
GPT-4oopenai· 128K ctx23$5.00cloudflare-ai-gateway$10.003.4×96%–
z-ai/glm-latestz-aiopen· 1.0M ctx5$2.65kilo$3.203.4×100%–
Nemotron 3.5 Lightning 30B A3Bnvidiaopen· 1M ctx9$0.150runinfra$0.2003.3×89%–
GPT-4o miniopenai· 128K ctx25$0.300cloudflare-ai-gateway$0.6003.3×96%–
tee/glm-5.2-thinkingteeopen· 1.0M ctx2$1.32nano-gpt$2.863.3×100%–
modelsdev/qwen2532binstructmodelsdevopen· 131K ctx2$0.861alibaba-cn$1.833.3×0%–
modelsdev/qwen2514binstructmodelsdevopen· 131K ctx2$0.431alibaba-cn$0.9153.2×0%–
Claude Sonnet 4.6anthropic· 1M ctx52$5.55xpersona$15.003.2×90%yes
Claude Fable 5anthropic· 1M ctx40$18.50xpersona$50.003.2×93%–
MiniMax-M3minimaxopen· 1.0M ctx46$0.900empiriolabs$1.203.2×83%yes
GPT-4.1 nanoopenai· 1.0M ctx25$0.260sap-ai-core$0.4003.1×92%–
Qwen3 VL 235B A22B Instructalibabaopen· 131K ctx20$0.880llmgateway$1.503.0×35%–
Gemini Flash Latestgoogle· 1.0M ctx9$3.00orcarouter$3.753.0×100%–
openai/gpt-chat-latestopenai· 1.1M ctx5$10.00nano-gpt$30.003.0×100%–
openai/gpt-5.6-sol-proopenai· 1.1M ctx3$10.00nano-gpt$10.003.0×67%yes
modelsdev/hermes3llama31405bmodelsdevopen· 128K ctx3$1.00kilo$1.003.0×0%–
meta-llama/llama-prompt-guard-2-22mmeta-llamaopen· 512 ctx2$0.010helicone$0.0203.0×0%–
Mistral Small (latest)mistralopen· 256K ctx5$0.200helicone$0.6003.0×40%–
Seed Characterbytedance-seed· 256K ctx2$0.297volcengine$0.5903.0×100%yes
modelsdev/deepseekr10528modelsdev· 128K ctx13$1.70nano-gpt$2.292.9×54%–
modelsdev/gemma4modelsdevopen· 262K ctx2$1.71greenpt$3.352.9×50%–
Gemma 3 12B ITgoogleopen· 1.0M ctx12$0.100novita-ai$0.1502.9×8%–
modelsdev/deepseekr1distillqwen14bmodelsdev· 33K ctx2$0.150novita-ai$0.2902.9×0%–
modelsdev/llama4scout17b16einstructmodelsdevopen· 128K ctx5$0.300deepinfra$0.7802.8×0%–
DeepSeek-V3deepseekopen· 164K ctx14$0.890edenai$1.132.8×21%–
GLM-4.5zhipuaiopen· 131K ctx20$0.800submodel$2.202.8×75%yes
Llama 4 Maverick 17B Instructmetaopen· 131K ctx9$0.590abacus$0.9702.7×0%–
Gemma 4 26B A4B ITgoogleopen· 262K ctx21$0.220kilo$0.4002.7×29%–
qwen/qwen2.5-vl-32b-instructqwenopen· 32K ctx2$0.220io-net$0.4102.7×50%–
modelsdev/qwen3vl235ba22bmodelsdevopen· 131K ctx4$1.15alibaba-cn$2.352.7×50%–
tee/gemma4-31bteeopen· 262K ctx4$0.370chutes$0.4102.7×100%–
modelsdev/qwen330ba3binstruct2507modelsdev· 128K ctx9$0.193openrouter$0.3002.7×33%–
moonshotai/kimi-latestmoonshotaiopen· 1.0M ctx6$8.50kilo$9.472.6×100%–
GLM-4.6zhipuaiopen· 205K ctx27$1.14302ai$2.202.6×81%yes
GLM-5.1zhipuaiopen· 203K ctx61$2.15crof$4.402.6×89%yes
Gemma 3 4B ITgoogleopen· 8K ctx10$0.080amazon-bedrock$0.1002.5×10%–
GPT-5openai· 400K ctx30$8.00ofox$10.002.5×97%–
LongCat-2.0meituan· 1.0M ctx5$1.20opencode-go$2.952.5×100%–
Qwen3.6 Plusalibaba· 1M ctx27$1.50kilo$3.002.5×81%yes
Ministral 3 14Bmistralopen· 262K ctx4$0.200digitalocean$0.4292.5×25%–
Gemini 3.1 Pro Previewgoogle· 1.0M ctx36$6.00kilo$12.002.5×94%yes
Gemini 3.5 Flash Litegoogle· 1.0M ctx28$1.25kilo$2.502.5×96%–
GPT-5.1 Codex Maxopenai· 400K ctx13$8.00ofox$10.002.5×85%–
GPT-5.3 Codexopenai· 400K ctx30$11.20ofox$14.002.5×93%–
GPT-5.4 miniopenai· 400K ctx37$3.60ofox$4.502.5×92%–
QwQ 32Balibabaopen· 41K ctx3$0.400abacus$0.8612.5×0%–
Ministral 3Bmistralopen· 128K ctx8$0.040vercel$0.0702.5×25%–
Sonarperplexity· 130K ctx10$1.00sap-ai-core$1.002.5×10%–
nousresearch/hermes-4-405bnousresearchopen· 128K ctx5$1.20nano-gpt$3.002.5×40%–
Llama-3.1-70B-Instructmetaopen· 131K ctx9$0.400openrouter$0.7202.5×11%–
meta-llama/llama-4-maverick-17b-128e-instruct-fp8meta-llamaopen· 131K ctx6$0.600io-net$0.9252.5×17%–
Hy3tencentopen· 262K ctx9$0.260nano-gpt$0.5802.5×89%–
DeepSeek-V3.1deepseekopen· 128K ctx20$0.700nano-gpt$1.002.5×45%–
Claude Opus 5anthropic· 1M ctx41$25.00aihubmix$25.002.4×95%–
Solar Pro 3upstage· 131K ctx5$0.250upstage$0.6002.4×60%–
Grok 4.20 (Reasoning)xai· 1M ctx5$2.50xai$2.502.4×100%yes
Grok 4.20 (Non-Reasoning)xai· 1M ctx5$2.50merge-gateway$2.502.4×100%yes
Solar Pro 2upstage· 66K ctx2$0.250upstage$0.4252.4×0%–
modelsdev/qwen3vl30ba3bthinkingmodelsdev· 128K ctx5$1.00novita-ai$1.002.4×0%–
x-ai/grok-4.20-multi-agentx-ai· 2M ctx6$2.50kilo$2.672.4×100%yes
vertex-openai/grok-4-20-reasoningvertex-openai· 2M ctx5$2.50llmgateway-providers$2.502.4×80%yes
vertex-openai/grok-4-20-non-reasoningvertex-openai· 2M ctx5$2.50vercel$2.502.4×80%yes
Inklingthinkingmachinesopen· 1.0M ctx24$4.05edenai$4.052.3×88%–
Hy3 previewtencentopen· 262K ctx4$0.260siliconflow$0.5862.3×100%–
Qwen Turboalibaba· 1M ctx5$0.087alibaba-cn$0.0902.3×60%–
Phi-4-minimicrosoftopen· 128K ctx3$0.300azure$0.3002.3×33%–
Qwen3 Coder Nextalibabaopen· 262K ctx18$0.675llmgateway-providers$1.352.2×28%–
Mistral Large 3mistralopen· 262K ctx11$1.50pioneer$1.502.2×45%–
Qwen-VL Plusalibaba· 131K ctx3$0.287alibaba-cn$0.6302.2×33%–
GPT-4.1 miniopenai· 1.0M ctx27$1.28ofox$1.602.2×96%–
GPT-4.1openai· 1.0M ctx27$6.40ofox$8.002.2×96%–
Qwen3.8 Omni Flashalibaba· 1M ctx4$0.430crossmodel$0.4702.2×100%–
modelsdev/qwen3omniflashrealtimemodelsdev· 66K ctx2$0.918alibaba-cn$1.452.2×0%–
GLM-4.6Vzhipuaiopen· 131K ctx13$0.430302ai$0.9002.1×69%yes
GLM-4.5Vzhipuaiopen· 66K ctx14$0.860302ai$1.802.1×64%–
Qwen3 VL 235B A22B Thinkingalibaba· 262K ctx13$2.87merge-gateway$3.952.1×15%–
Qwen3.6 Max Previewalibaba· 262K ctx14$6.16kilo$7.802.1×79%yes
Claude Sonnet 5anthropic· 1M ctx42$7.20unorouter$10.002.1×93%–
z-ai/glm-5z-aiopen· 205K ctx59$1.92gmicloud$3.202.1×81%yes
tee/glm-5.1-thinkingteeopen· 203K ctx2$2.60nano-gpt$3.922.0×100%–
Gemini 3.1 Flash Litegoogle· 1.0M ctx12$0.750kilo$1.502.0×92%–
Gemini 3.8 Flashgoogle· 1.0M ctx24$3.75llmgateway-providers$3.752.0×96%–
Gemini 3.7 Flashgoogle· 1.0M ctx28$3.75opper$3.752.0×96%–
o4-miniopenai· 200K ctx21$4.00poe$4.402.0×90%–
Gemini 3.1 Flash Lite Previewgoogle· 1.0M ctx6$0.750kilo$1.502.0×83%–
o3-proopenai· 200K ctx8$40.00abacus$80.002.0×0%–
Gemma 4 E4B ITgoogleopen· 33K ctx2$0.100deepinfra$0.1502.0×50%–
Nano Banana Pro Previewgoogle· 66K ctx4$6.00kilo$12.002.0×75%–
thedrummer/cydonia-24b-v2thedrummeropen· 33K ctx2$0.121nano-gpt$0.1812.0×100%–
azure/gpt-4-turboazure· 128K ctx8$30.00azure$30.002.0×0%–
openai/gpt-5.1-thinking-fastopenai· 400K ctx2$10.00vercel$15.002.0×100%–
modelsdev/devstral2123binstruct2512modelsdevopen· 200K ctx2$1.40nano-gpt$2.072.0×50%–
DeepSeek V3 0324deepseekopen· 164K ctx11$0.770nano-gpt$1.001.9×45%–
Gemini Omni Flash Previewgoogle· 1M ctx2$9.00vercel$13.251.9×0%–
qwen/qwen3-vl-30b-a3b-instructqwen· 262K ctx9$0.520kilo$0.7001.9×0%–
Pixtral 12Bmistralopen· 128K ctx3$0.150pioneer$0.1501.9×33%–
Seed 2.0 Minibytedance-seed· 262K ctx9$0.297volcengine$0.4001.9×67%yes
modelsdev/doubaoseed20codemodelsdev· 256K ctx2$2.41aihubmix$3.451.9×50%yes
undi95/remm-slerp-l2-13bundi95open· 6K ctx3$0.650kilo$0.6501.9×33%–
modelsdev/qwen3vl30ba3bmodelsdevopen· 131K ctx2$0.431alibaba-cn$0.6161.9×0%–
Qwen Flashalibaba· 1M ctx8$0.216merge-gateway$0.3101.9×63%–
Mistral Small 3.1 24Bmistralopen· 131K ctx3$0.300nano-gpt$0.3181.9×33%–
modelsdev/qwen330ba3bthinking2507modelsdev· 126K ctx3$1.30clarifai$2.401.8×0%–
Kimi K2.5moonshotai· 256K ctx45$1.90nano-gpt$3.001.8×69%–
Llama-Guard-3-8Bmetaopen· 131K ctx3$0.030edenai$0.0301.8×0%–
modelsdev/mistral7binstruct3modelsdevopen· 33K ctx3$0.110ovhcloud$0.1111.8×33%–
Seed 2.1 Turbobytedance-seed· 262K ctx8$1.77ofox$2.501.8×75%–
modelsdev/minimaxm1modelsdev· 1M ctx4$1.25302ai$1.771.8×25%–
modelsdev/qwenmtmodelsdev· 16K ctx2$0.280alibaba-cn$0.3851.7×0%–
nousresearch/hermes-3-llama-3.1-70bnousresearchopen· 66K ctx3$0.408nano-gpt$0.7001.7×33%–
Nova 2 Liteamazon· 1M ctx10$2.50amazon-bedrock$2.751.7×70%–
anthracite-org/magnum-v4-72banthracite-orgopen· 16K ctx4$2.99nano-gpt$4.001.7×50%–
Gemini Flash-Lite Latestgoogle· 1.0M ctx5$1.50google-vertex$1.501.7×100%–
Laguna XS 2.1poolsideopen· 262K ctx2$0.120openrouter$0.1601.7×100%–
ibm-granite/granite-4.2-8bibm-graniteopen· 131K ctx4$0.150nano-gpt$0.2001.7×100%–
modelsdev/gpt56modelsdev· 372K ctx2$12.00xpersona$16.001.7×100%yes
moonshotai/kimi-k2-instruct-0905moonshotaiopen· 262K ctx3$1.80nano-gpt$1.901.7×67%–
modelsdev/kimik2instructmodelsdevopen· 131K ctx6$1.80nano-gpt$2.051.7×50%–
mistralai/ministral-14b-2512mistralaiopen· 262K ctx4$0.200openrouter$0.2201.6×75%–
mistralai/mistral-small-3.2-24b-instructmistralai· 128K ctx4$0.250kilo$0.2501.6×25%–
Seed 2.0 Probytedance-seed· 256K ctx6$2.37volcengine$3.001.6×83%yes
GPT-5.4 Proopenai· 1.1M ctx20$144ofox$1801.6×15%yes
qwen/qwen3.8-maxqwen· 1M ctx27$4.80vancine$6.001.6×89%–
Mistral Small 4mistralopen· 256K ctx12$0.600kilo$0.6001.6×67%–
alibaba/qwen3-max-thinkingalibabaopen· 256K ctx3$3.90openrouter$3.901.5×33%yes
meta-llama/llama-4-scoutmeta-llamaopen· 328K ctx4$0.300helicone$0.3001.5×25%–
MiniMax-M2minimaxopen· 197K ctx17$1.00llmgateway-providers$1.201.5×47%–
Mistral Small 3.2mistralopen· 128K ctx4$0.300llmgateway-providers$0.3001.5×0%–
Qwen3.6 Flashalibaba· 992K ctx20$0.990merge-gateway$1.131.5×55%yes
Claude Opus 4.7anthropic· 1M ctx50$20.00kilo$25.001.5×98%yes
Claude Opus 4.6anthropic· 1M ctx45$20.00kilo$25.001.5×96%yes
Magistral Medium (latest)mistral· 128K ctx4$5.00pioneer$5.001.5×50%–
sakana/fugu-ultra-v2.0sakana· 1M ctx4$30.00llmgateway-providers$30.001.5×100%yes
GPT OSS Safeguard 20Bopenaiopen· 131K ctx10$0.200amazon-bedrock$0.2501.5×40%–
qwen/qwen3-vl-8b-instructqwen· 262K ctx5$0.455openrouter$0.5001.5×0%–
Qwen3.8 Flashalibaba· 992K ctx23$0.380vancine$0.4701.5×96%–
modelsdev/qwen25vl7binstructmodelsdev· 128K ctx2$0.717alibaba-cn$0.8831.5×0%–
nvidia/nvidia-nemotron-3-ultra-550b-a55bnvidiaopen· 203K ctx3$2.15wandb$2.401.5×100%–
meta-llama/llama-4-maverickmeta-llamaopen· 1.0M ctx5$0.600nano-gpt$0.6521.4×20%–
Claude Sonnet 4.5 (latest)anthropic· 200K ctx25$13.00poe$15.001.4×100%yes
qwen/qwen3-vl-32b-instructqwen· 262K ctx4$0.416openrouter$0.5081.4×0%–
Hy4 previewtencentopen· 1.0M ctx11$2.00vancine$2.501.4×100%–
Gemini 2.5 Progoogle· 1.0M ctx33$7.00poe$10.001.4×91%yes
Gemini 2.0 Flash-Litegoogle· 990K ctx2$0.210poe$0.2551.4×0%–
Claude Opus 4.5 (latest)anthropic· 200K ctx22$21.00poe$25.001.4×100%–
GPT-3.5-turboopenai· 16K ctx14$1.40poe$1.501.4×21%–
deepseek/deepseek-v3.1-terminusdeepseek· 128K ctx11$0.700nano-gpt$1.001.4×55%–
openai/chatgpt-4o-latestopenai· 128K ctx2$14.00poe$17.001.4×50%–
Command R+cohereopen· 128K ctx6$10.00edenai$10.001.4×17%–
qwen/qwen-longqwen· 10M ctx2$0.287alibaba-cn$0.3471.4×50%–
Seed 2.0 Codebytedance-seed· 262K ctx10$2.37volcengine$3.001.4×70%yes
modelsdev/qwen3235ba22bthinkingmodelsdev· 262K ctx2$2.90helicone$3.451.4×0%–
GPT-5 Nanoopenai· 400K ctx29$0.320ofox$0.4001.4×97%–
GPT-5.1openai· 400K ctx31$8.00ofox$10.001.4×94%–
GPT-5.5 Proopenai· 1.1M ctx19$164poe$1801.4×11%yes
Seed 1.6 Flashbytedance-seed· 256K ctx6$0.220ofox$0.3001.4×67%yes
Claude Haiku 4.5 (latest)anthropic· 200K ctx30$3.70xpersona$5.001.4×97%–
Nemotron 3.5 Content Safetynvidiaopen· 131K ctx3$0.150nano-gpt$0.2001.3×33%–
Voxtral Small (latest)mistralopen· 32K ctx3$0.300mistral$0.3001.3×33%–
Sarvam 105Bsarvamopen· 131K ctx2$0.160fastrouter$0.1861.3×50%–
deepseek/deepseek-v3.2-expdeepseek· 128K ctx8$0.330zenmux$0.4101.3×13%–
modelsdev/nemotron3ultra550bmodelsdev· 131K ctx2$1.70digitalocean$1.951.3×50%–
qwen/qwen3-235b-a22b-2507qwenopen· 131K ctx2$0.464tensorx$0.5311.3×50%–
modelsdev/openaigptoss120bmodelsdevopen· 128K ctx2$0.300venice$0.3431.3×50%–
moonshotai/kimi-k2-0905moonshotai· 256K ctx7$2.00helicone$2.501.3×29%–
Qwen3.8 2.4T A95Balibabaopen· 262K ctx18$5.95iteracompute$6.001.3×94%–
Seed 2.1 Probytedance-seed· 256K ctx2$3.54ofox$3.991.3×100%–
microsoft/wizardlm-2-8x22bmicrosoftopen· 66K ctx4$0.493nano-gpt$0.6201.3×25%–
GPT-5.1 Codex miniopenai· 400K ctx18$1.60ofox$2.001.3×94%–
GPT-6 Astraopenai· 1.1M ctx10$40.00ofox$50.001.3×100%yes
GPT-5.4 nanoopenai· 400K ctx32$1.00ofox$1.251.3×91%–
GLM-4.7-Flashzhipuaiopen· 203K ctx27$0.400cloudflare-workers-ai$0.4001.3×59%–
Gemini Deep Research Previewgoogle· 1.0M ctx2$9.60poe$10.801.3×50%yes
Gemma 4 E2B ITgoogleopen· 33K ctx2$0.080amazon-bedrock$0.0901.3×50%–
modelsdev/zaiglm5modelsdevopen· 200K ctx2$3.20amazon-bedrock$3.601.3×50%–
vertex-openai/kimi-k2-thinkingvertex-openaiopen· 262K ctx7$2.00helicone$2.501.3×57%–
deepseek/deepseek-chat-v3.1deepseekopen· 164K ctx2$0.800tensorx$0.9001.3×100%–
modelsdev/mercury2modelsdev· 128K ctx6$0.750openrouter$0.7501.3×100%–
modelsdev/nvidianemotron3nano30ba3bmodelsdevopen· 128K ctx2$0.240cortecs$0.2701.3×0%–
mistralai/ministral-3b-2512mistralaiopen· 131K ctx3$0.100openrouter$0.1001.2×100%–
mistralai/codestral-2508mistralai· 256K ctx4$0.900nano-gpt$0.9001.2×100%–
gryphe/mythomax-l2-13bgrypheopen· 4K ctx4$0.090novita-ai$0.1051.2×25%–
Llama 4 Scout 17B Instructmetaopen· 8K ctx6$0.590llmgateway$0.6601.2×0%–
Claude Fable 5.1anthropic· 1M ctx29$50.00amazon-bedrock$50.001.2×97%–
Inkling Smallthinkingmachinesopen· 524K ctx11$1.20nano-gpt$1.201.2×82%–
DeepSeek V4 Pro 0423deepseekopen· 1M ctx2$3.30merge-gateway$3.631.2×50%–
Devstral 2mistralopen· 262K ctx9$2.00edenai$2.001.2×56%–
mistralai/ministral-8b-2512mistralaiopen· 262K ctx6$0.150llmgateway-providers$0.1501.2×67%–
GPT-5.1 Codexopenai· 400K ctx19$8.50opencode$10.001.2×89%–
GPT-5-Codexopenai· 400K ctx10$8.50opencode$10.001.2×80%–
Claude Haiku 3.5anthropic· 189K ctx2$3.40poe$3.701.2×100%–
Claude Opus 4.1 (latest)anthropic· 200K ctx13$64.00poe$75.001.2×100%–
anthropic/claude-opus-4anthropic· 193K ctx7$64.00poe$75.001.2×100%–
Qwen3.8 Max 0902alibaba· 992K ctx8$5.14ofox$6.001.2×100%–
modelsdev/qwenomnirealtimemodelsdev· 33K ctx2$0.918alibaba-cn$0.9941.2×0%–
GPT OSS Safeguard 120Bopenaiopen· 131K ctx5$0.600merge-gateway$0.6001.2×0%–
Nova Liteamazon· 300K ctx12$0.238nano-gpt$0.2401.2×75%–
modelsdev/qwen25omni7bmodelsdevopen· 33K ctx2$0.345alibaba-cn$0.3721.2×0%–
Claude Sonnet 3.7anthropic· 200K ctx4$13.00poe$15.001.2×75%–
anthropic/claude-sonnet-4anthropic· 983K ctx10$13.00poe$15.001.2×100%yes
Nova Proamazon· 300K ctx11$3.20nano-gpt$3.201.2×73%–
qwen/qwen3-embedding-8bqwenopen· 41K ctx5$0.100scaleway$0.1081.1×0%–
Nova Microamazon· 128K ctx10$0.140edenai$0.1401.1×70%–
Claude Haiku 3anthropic· 189K ctx3$1.10poe$1.251.1×100%–
Grok 4.5xai· 500K ctx26$6.00kilo$6.001.1×92%yes
Grok 4.6xai· 500K ctx32$6.00abacus$6.001.1×97%yes
Grok 4.3xai· 1M ctx30$2.50auriko$2.501.1×93%yes
modelsdev/mistralsmall3224binstruct2506modelsdevopen· 128K ctx4$0.310ovhcloud$0.3211.1×0%–
Trinity Large Thinkingarcee-aiopen· 262K ctx5$0.800openrouter$0.8001.1×80%–
GPT-5.2 Proopenai· 400K ctx12$150poe$1681.1×0%–
Qwen3-VL Plusalibaba· 262K ctx6$1.43alibaba-cn$1.601.1×50%–
modelsdev/pixtral12b2409modelsdevopen· 128K ctx2$0.200scaleway$0.2121.1×0%–
modelsdev/mistralsmall2503modelsdev· 128K ctx3$0.300azure-cognitive-services$0.3001.1×0%–
Grok 4.1 Fast (Reasoning)xai· 2M ctx9$0.450jiekou$0.5001.1×67%–
Laguna S 2.1poolsideopen· 1.0M ctx5$0.180openrouter$0.2001.1×100%–
GPT-5 Proopenai· 400K ctx16$108jiekou$1201.1×6%–
GPT-4 Turboopenai· 128K ctx10$27.00poe$30.001.1×0%–
o1openai· 200K ctx15$54.00poe$60.001.1×93%–
o1-proopenai· 200K ctx6$540poe$6001.1×17%–
Claude Opus 4.1anthropic· 200K ctx7$67.50jiekou$75.001.1×57%–
Claude Opus 4anthropic· 200K ctx3$67.50jiekou$75.001.1×33%–
Claude Sonnet 4anthropic· 200K ctx3$13.50jiekou$15.001.1×33%–
GPT-5 Chat (latest)openai· 400K ctx5$9.00jiekou$10.001.1×80%–
x-ai/grok-4-fast-reasoningx-ai· 2M ctx4$0.450jiekou$0.5001.1×50%–
modelsdev/gemini25pro0605modelsdev· 1.0M ctx2$9.00jiekou$9.501.1×50%–
x-ai/grok-4.1-fast-non-reasoningx-ai· 2M ctx11$0.450jiekou$0.5001.1×73%–
x-ai/grok-4-fast-non-reasoningx-ai· 2M ctx5$0.450jiekou$0.5001.1×40%–
modelsdev/gemini25flashlite092025modelsdev· 1.0M ctx2$0.360jiekou$0.3801.1×0%–
modelsdev/gpt35instructmodelsdev· 4K ctx5$1.80poe$2.001.1×0%–
modelsdev/grok40709modelsdev· 256K ctx2$13.50jiekou$14.251.1×0%–
x-ai/grok-code-fast-1x-ai· 256K ctx6$1.35jiekou$1.501.1×67%–
minimaxai/minimax-m1-80kminimaxaiopen· 10.2M ctx3$2.20jiekou$2.201.1×33%–
o3-miniopenai· 200K ctx19$4.00poe$4.401.1×89%–
modelsdev/openaigpt6astramodelsdev· 1.1M ctx3$50.00venice$50.001.1×100%yes
thedrummer/cydonia-24b-v4.1thedrummeropen· 131K ctx3$0.500kilo$0.5001.1×100%–
openai/o3-mini-highopenai· 200K ctx4$4.00poe$4.401.1×75%–
modelsdev/devstralmodelsdev· 256K ctx2$2.00mistral$2.101.1×50%–
modelsdev/gemini31flashlite@eumodelsdev· 1.0M ctx2$1.50edenai$1.571.1×100%–
novita/hy3novitaopen· 262K ctx8$0.528llmgateway$0.5801.1×88%–
modelsdev/qwen3asrflashmodelsdev· 53K ctx2$0.032alibaba-cn$0.0341.1×0%–
modelsdev/gemini31flashlitemodelsdev· 1.0M ctx19$1.50abacus$1.501.1×89%–
Apertus 70Bswiss-aiopen· 524K ctx2$2.23cortecs$2.321.1×0%–
GPT-Image-2openai4$30.00openai$30.001.1×100%–
modelsdev/doubaoseed20promodelsdev· 256K ctx2$2.24zenmux$2.331.1×100%yes
MiMo-V2-Flashxiaomiopen· 262K ctx7$0.280xiaomi$0.3001.1×71%–
Qwen3.8 Flash Nextalibabaopen· 1.0M ctx5$0.470amd$0.5001.1×100%–
modelsdev/qvqmaxmodelsdevopen· 128K ctx3$4.59alibaba-cn$4.801.0×33%–
modelsdev/ernie50thinkingmodelsdev· 128K ctx2$3.37zenmux$3.441.0×50%–
Step 3.7 Flashstepfunopen· 262K ctx13$1.11stepfun$1.151.0×77%–
Seed 1.6 Visionbytedance-seed· 256K ctx2$1.15ofox$1.171.0×100%yes
modelsdev/ling1tmodelsdevopen· 128K ctx2$2.24zenmux$2.271.0×50%–
modelsdev/ring1tmodelsdevopen· 128K ctx2$2.24zenmux$2.271.0×50%–
minimax/minimax-01minimax· 1.0M ctx3$1.10kilo$1.101.0×33%–
tee/llama3-3-70bteeopen· 128K ctx2$2.75nano-gpt$2.771.0×50%–
mistralai/mistral-sabamistralaiopen· 33K ctx3$0.595nano-gpt$0.6001.0×100%–
Seed Evolvingbytedance-seed· 256K ctx2$4.42ofox$4.441.0×100%–
MiniMax-M2 Herminimax· 66K ctx4$1.20ofox$1.201.0×75%–
modelsdev/claude45sonnetmodelsdev· 200K ctx2$14.95cortecs$14.971.0×100%–
Step 3.5 Flashstepfunopen· 262K ctx16$0.300stepfun$0.3001.0×63%–
Command A Reasoningcohereopen· 256K ctx2$10.00nano-gpt$10.001.0×50%–
Paretounbiased· 262K ctx4$7.50kilo$7.501.0×100%–
Ministral 14Bmistral· 256K ctx5$0.200nano-gpt$0.2001.0×60%–
Mistral Medium 3mistral· 128K ctx7$2.00nano-gpt$2.001.0×43%–
Gemini 3.1 Pro Preview Custom Toolsgoogle· 1.0M ctx10$12.00merge-gateway$12.001.0×100%yes
Muse Spark 1.3meta· 1.0M ctx9$4.25opencode$4.251.0×100%–
Muse Spark 1.2meta· 1M ctx13$4.25empiriolabs$4.251.0×100%–
Muse Spark 1.1meta· 1M ctx12$4.25orcarouter$4.251.0×100%–
Fugu Ultrasakana· 1M ctx15$30.00sakana$30.001.0×93%yes
GPT-4o (2024-08-06)openai· 128K ctx7$10.00merge-gateway$10.001.0×100%–
GPT-4o (2024-11-20)openai· 128K ctx8$10.00orcarouter$10.001.0×88%–
Nemotron Nano 12B v2 VLnvidiaopen· 128K ctx2$0.600vercel$0.6001.0×0%–
GPT-Image-1openai· 128K ctx3$40.00azure$40.001.0×100%–
Whisper Large v3 Turboopenaiopen· 448 ctx3$0.0023privatemode-ai$0.00231.0×0%–
Qwen3.8 Maxalibaba· 1M ctx3$6.00cloudflare-ai-gateway$6.001.0×100%–
GPT-4openai· 8K ctx6$60.00edenai$60.001.0×0%–
Codestral (latest)mistralopen· 33K ctx7$0.900vercel$0.9001.0×29%–
Sonar Reasoning Properplexity· 128K ctx8$8.00edenai$8.001.0×0%–
Sonar Properplexity· 200K ctx9$15.00openrouter$15.001.0×0%–
GPT-4o (2024-05-13)openai· 128K ctx5$15.00kilo$15.001.0×0%–
GPT-5.1 Chatopenai· 128K ctx6$10.00zenmux$10.001.0×67%–
GPT-5.2 Chatopenai· 128K ctx7$14.00302ai$14.001.0×71%–
Sonar Deep Researchperplexity· 128K ctx7$8.00helicone$8.001.0×0%–
Claude Mythos 5anthropic· 1M ctx2$50.00azure-cognitive-services$50.001.0×100%–
Command Acohereopen· 256K ctx7$10.00azure-cognitive-services$10.001.0×0%–
Gemini 2.5 Flash TTSgoogle· 33K ctx2$10.00google$10.001.0×0%–
Gemini 2.5 Pro TTSgoogle· 33K ctx2$20.00google-vertex$20.001.0×0%–
GPT-Image-1.5openai· 128K ctx2$32.00vercel$32.001.0×100%–
GPT-5.3 Codex Sparkopenai· 128K ctx2$14.00openai$14.001.0×100%–
Mistral Large 2.1mistralopen· 131K ctx3$6.00merge-gateway$6.001.0×0%–
Gemini 2.5 Computer Use Previewgoogle· 128K ctx2$10.00google$10.001.0×0%yes
Nemotron Nano 9B v2nvidiaopen· 131K ctx2$0.230merge-gateway$0.2301.0×0%–
Nano Banana Progoogle· 66K ctx5$12.00merge-gateway$12.001.0×80%–
Nano Banana 2google· 131K ctx5$3.00openrouter$3.001.0×20%–
Nano Banana 2 Previewgoogle· 131K ctx5$3.00kilo$3.001.0×20%–
Sakana Namazusakana· 262K ctx5$4.00sakana$4.001.0×100%–
Ling 3.0 Flash Fininclusionai· 256K ctx2$0.180kilo$0.1801.0×100%–
GPT Image 2.5 Sunburstopenai2$30.00azure$30.001.0×100%–
GPT Image 2.5 Flareopenai2$30.00vercel$30.001.0×100%–
GPT-Realtime-2.1openai· 128K ctx2$24.00openai$24.001.0×100%–
GPT-5.3 Chat (latest)openai· 128K ctx6$14.00merge-gateway$14.001.0×50%–
Palmyra X5writer· 1.0M ctx5$6.00amazon-bedrock$6.001.0×0%–
Nova Premieramazon· 1M ctx3$12.50amazon-bedrock$12.501.0×100%–
Grok Build 0.1xai· 256K ctx18$2.00requesty$2.001.0×94%yes
Command R7Bcohereopen· 128K ctx5$0.150kilo$0.1501.0×0%–
Command Rcohereopen· 128K ctx5$0.600merge-gateway$0.6001.0×0%–
Palmyra X4writer· 128K ctx3$10.00amazon-bedrock$10.001.0×0%–
Voxtral Mini 3B 2507mistralopen· 33K ctx3$0.040edenai$0.0401.0×0%–
Devstral Smallmistralopen2$0.300merge-gateway$0.3001.0×0%–
Devstral Mediummistral· 128K ctx2$2.00merge-gateway$2.001.0×0%–
Devstral 2 (latest)mistralopen· 262K ctx3$2.00mistral$2.001.0×0%–
Pixtral Large (latest)mistralopen· 128K ctx2$6.00mistral$6.001.0×0%–
Magistral Smallmistralopen· 128K ctx3$1.50pioneer$1.501.0×67%–
Gemma-SEA-LION-v4-27B-ITaisingaporeopen· 128K ctx2$0.555cloudflare-workers-ai$0.5551.0×0%–
Granite-4.0-H-Microibm· 131K ctx3$0.112cloudflare-workers-ai$0.1121.0×0%–
openai/gpt-luna-latestopenai· 1.1M ctx3$1.20kilo$1.201.0×100%yes
qwen/qwen-plus-2025-07-28qwen· 1M ctx2$0.780kilo$0.7801.0×0%yes
meta/muse-spark-1.2-contributormeta· 1M ctx9$0.200llmtr$0.2001.0×89%–
x-ai/grok-latestx-ai· 500K ctx4$6.00openrouter$6.001.0×100%yes
modelsdev/phi4reasoningmodelsdevopen· 32K ctx2$0.500azure$0.5001.0×0%–
modelsdev/claudeopus4@20250514modelsdev· 200K ctx2$75.00google-vertex-anthropic$75.001.0×100%–
modelsdev/morphmodelsdev· 16K ctx4$1.20vercel$1.201.0×0%–
modelsdev/moonshotkimik2thinkingmodelsdevopen· 262K ctx2$2.50amazon-bedrock$2.501.0×0%–
modelsdev/gpt350125modelsdev· 16K ctx2$1.50azure$1.501.0×0%–
modelsdev/anthropicclaude35sonnetmodelsdev· 200K ctx2$15.00digitalocean$15.001.0×100%–
zai/glm-5.3-flashxzaiopen· 1M ctx6$1.25vercel$1.251.0×100%–
tencent/hy-mt2-1.8btencent· 8K ctx2$0.177openrouter$0.1771.0×0%–
modelsdev/musespark13contributormodelsdev· 1.0M ctx9$0.200vercel$0.2001.0×89%–
sao10k/l3.1-euryale-70bsao10k· 131K ctx2$0.850kilo$0.8501.0×0%–
modelsdev/anthropicclaude46sonnetmodelsdev· 200K ctx2$15.00digitalocean$15.001.0×100%yes
mancer/weavermancer· 8K ctx2$0.750kilo$0.7501.0×0%–
modelsdev/phi4multimodalmodelsdevopen· 128K ctx2$0.320azure-cognitive-services$0.3201.0×0%–
modelsdev/glm52shortflexmodelsdevopen· 200K ctx2$2.92neuralwatt$2.921.0×100%–
modelsdev/claude35haikumodelsdev· 200K ctx2$4.00zenmux$4.001.0×100%–
alibaba/qwen-coder-plusalibaba· 131K ctx2$1.00llmgateway$1.001.0×0%–
modelsdev/gpt6astramodelsdev· 1.1M ctx11$50.00azure$50.001.0×91%yes
baidu/ernie-4.5-21b-a3b-thinkingbaiduopen· 131K ctx2$0.280novita-ai$0.2801.0×0%–
openai/gpt-4o-transcribeopenai· 16K ctx3$10.00vercel$10.001.0×0%–
modelsdev/codexminimodelsdev· 200K ctx2$6.00azure-cognitive-services$6.001.0×100%–
google/gemma-2-27b-itgoogle· 8K ctx2$0.650kilo$0.6501.0×0%–
perceptron/perceptron-mk1perceptron· 33K ctx3$1.50kilo$1.501.0×33%–
tencent/hy-mt2-7btencent· 8K ctx2$0.295openrouter$0.2951.0×0%–
tencent/hy-mt2-30b-a3btencent· 8K ctx2$0.295kilo$0.2951.0×0%–
inclusionai/ling-2.6-1tinclusionaiopen· 262K ctx2$2.50requesty$2.501.0×50%–
modelsdev/llama3370bversatilemodelsdevopen· 131K ctx3$0.790groq$0.7901.0×0%–
openai/gpt-3.5-turbo-16kopenai· 16K ctx2$4.00openrouter$4.001.0×0%–
modelsdev/claudesonnet45@20250929modelsdev· 200K ctx2$15.00google-vertex$15.001.0×100%–
rekaai/reka-flash-3rekaai· 66K ctx2$0.200kilo$0.2001.0×0%–
abliteration-ai/abliterated-modelabliteration-ai· 262K ctx2$3.00abliteration-ai$3.001.0×100%–
openai/gpt-terra-latestopenai· 1.1M ctx3$12.00kilo$12.001.0×100%yes
liquidai/lfm2-24b-a2bliquidaiopen· 33K ctx2$0.120pioneer$0.1201.0×50%–
modelsdev/claudeopus41@20250805modelsdev· 200K ctx2$75.00google-vertex-anthropic$75.001.0×100%–
sakana/fugu-ultra-v1.1sakana· 1M ctx2$30.00nano-gpt$30.001.0×100%yes
kwaipilot/kat-coder-prokwaipilotopen· 256K ctx5$1.20novita-ai$1.201.0×80%–
anthropic/claude-sonnet-4.6-thinkinganthropic· 1M ctx2$15.00nano-gpt$15.001.0×50%–
anthropic/claude-sonnet-latestanthropic· 1M ctx4$10.00kilo$10.001.0×100%–
modelsdev/anthropicclaude37sonnetmodelsdev· 200K ctx2$15.00sap-ai-core$15.001.0×100%–
modelsdev/phi4minireasoningmodelsdevopen· 128K ctx2$0.300azure-cognitive-services$0.3001.0×0%–
aion-labs/aion-rp-llama-3.1-8baion-labs· 33K ctx3$1.60nano-gpt$1.601.0×33%–
openai/o4-mini-highopenai· 200K ctx3$4.40openrouter$4.401.0×100%–
modelsdev/phi4reasoningplusmodelsdevopen· 32K ctx2$0.500azure-cognitive-services$0.5001.0×0%–
openai/gpt-5.6-luna-proopenai· 1.1M ctx3$1.20openrouter$1.201.0×67%yes
bytedance/ui-tars-1.5-7bbytedance· 128K ctx2$0.200openrouter$0.2001.0×100%–
minimax/minimax-text-01minimax· 1M ctx2$1.10llmgateway-providers$1.101.0×0%–
google/gemini-pro-latestgoogle· 1.0M ctx7$12.00edenai$12.001.0×100%yes
rekaai/reka-edgerekaai· 16K ctx2$0.100kilo$0.1001.0×0%–
inclusionai/ling-flash-2.0inclusionai· 131K ctx2$0.570siliconflow-cn$0.5701.0×0%–
qwen/qwen3.5-plus-20260420qwen· 1M ctx2$1.80openrouter$1.801.0×0%yes
inference-net/schematron-v2-turboinference-net· 128K ctx4$0.150nano-gpt$0.1501.0×100%–
qwen/qwen3.5-plus-02-15qwen· 1M ctx2$1.56kilo$1.561.0×0%yes
openai/gpt-latestopenai· 1.1M ctx2$50.00nano-gpt$50.001.0×100%yes
inference.net/llama-3.2-11b-instructinference.net· 128K ctx2$0.330llmgateway$0.3301.0×0%–
modelsdev/deepseek2specialemodelsdevopen· 128K ctx2$1.68azure-cognitive-services$1.681.0×0%–
modelsdev/llama318binstantmodelsdevopen· 131K ctx2$0.080groq$0.0801.0×0%–
inclusionai/ling-3.0-flash-vlinclusionaiopen· 262K ctx3$0.180openrouter$0.1801.0×100%–
mistralai/mistral-small-3.1-24b-instructmistralai· 128K ctx2$0.555cloudflare-workers-ai$0.5551.0×0%–
modelsdev/qwen3635bmodelsdevopen· 131K ctx2$1.15neuralwatt$1.151.0×100%–
baidu/ernie-4.5-300b-a47b-paddlebaiduopen· 131K ctx2$1.10jiekou$1.101.0×0%–
modelsdev/mercurycodersmallmodelsdev· 33K ctx2$1.00nano-gpt$1.001.0×50%–
aion-labs/aion-3.0-miniaion-labs· 131K ctx3$1.40openrouter$1.401.0×100%–
anthropic/claude-haiku-latestanthropic· 200K ctx3$5.00openrouter$5.001.0×100%–
openai/gpt-realtime-1.5openai2$16.00fastrouter$16.001.0×50%–
kwaipilot/kat-coder-pro-v2.5kwaipilot· 256K ctx2$2.96openrouter$2.961.0×100%–
mistralai/mistral-large-2407mistralai· 131K ctx2$6.00kilo$6.001.0×100%–
mistralai/mixtral-8x22b-instructmistralaiopen· 66K ctx2$6.00openrouter$6.001.0×100%–
modelsdev/claudesonnet4@20250514modelsdev· 200K ctx2$15.00google-vertex$15.001.0×100%–
openai/gpt-5-imageopenai· 400K ctx2$10.00kilo$10.001.0×100%–
modelsdev/anthropicclaude45haikumodelsdev· 200K ctx2$5.00sap-ai-core$5.001.0×100%–
xai/grok-4-20-beta-0309-reasoningxai· 2M ctx3$6.00302ai$6.001.0×67%yes
anthropic/claude-opus-latestanthropic· 1M ctx4$25.00edenai$25.001.0×100%–
abliteration-ai/abliterated-model-largeabliteration-ai· 1M ctx4$5.00nano-gpt$5.001.0×100%–
modelsdev/step216kmodelsdev· 16K ctx2$16.44stepfun$16.441.0×100%–
zai/glm-4-32b-0414-128kzai· 128K ctx2$0.100llmgateway-providers$0.1001.0×0%–
nousresearch/hermes-2-pro-llama-3-8bnousresearchopen· 8K ctx2$0.140helicone$0.1401.0×0%–
openai/gpt-5.4-image-2openai· 272K ctx2$15.00openrouter$15.001.0×100%–
openai/gpt-4o-mini-transcribeopenai· 16K ctx3$5.00vercel$5.001.0×0%–
openai/gpt-5-image-miniopenai· 400K ctx2$2.00openrouter$2.001.0×100%–
modelsdev/qwen3embedding06bmodelsdevopen· 8K ctx2$0.010digitalocean$0.0101.0×0%–
openai/gpt-audioopenai· 128K ctx2$10.00openrouter$10.001.0×0%–
thedrummer/skyfall-36b-v2thedrummeropen· 33K ctx3$0.800openrouter$0.8001.0×100%–
zai/glm-4.6v-flashxzai· 128K ctx2$0.400llmgateway-providers$0.4001.0×100%–
~openai/gpt-mini-latest~openai· 400K ctx3$4.50openrouter$4.501.0×100%–
tencent/hy-mt2-plustencent· 8K ctx3$0.295llmgateway-providers$0.2951.0×0%–
modelsdev/anthropicclaude45sonnetmodelsdev· 200K ctx2$15.00sap-ai-core$15.001.0×100%yes
alibaba/qwen3-vl-flashalibaba· 262K ctx2$0.400llmgateway-providers$0.4001.0×100%–
modelsdev/morphlargemodelsdev· 32K ctx4$1.90morph$1.901.0×0%–
anthropic/claude-fable-latestanthropic· 1M ctx4$50.00nano-gpt$50.001.0×100%–
bytedance-seed/seed-oss-36b-instructbytedance-seed· 262K ctx2$0.570siliconflow$0.5701.0×0%–
mistral/ministral-8bmistralopen· 128K ctx2$0.100mistral$0.1001.0×0%–
sao10k/l3.3-euryale-70bsao10k· 131K ctx2$0.750openrouter$0.7501.0×0%–
modelsdev/gpt351106modelsdev· 16K ctx2$2.00azure$2.001.0×0%–
minimax/minimax-latestminimaxopen· 512K ctx2$1.20fireworks-ai$1.201.0×100%–
openai/gpt-audio-miniopenai· 128K ctx2$2.40kilo$2.401.0×0%–
modelsdev/step132kmodelsdev· 33K ctx2$9.59stepfun-ai$9.591.0×100%–
modelsdev/gpt4visionmodelsdev· 128K ctx2$30.00azure$30.001.0×0%–
thedrummer/unslopnemo-12bthedrummer· 33K ctx2$0.400kilo$0.4001.0×0%–
novita/llama-3-70b-instructnovita· 8K ctx3$0.740llmgateway-providers$0.7401.0×0%–
relace/relace-apply-3relace· 256K ctx2$1.25openrouter$1.251.0×0%–
inference-net/schematron-v2-smallinference-net· 128K ctx4$0.230openrouter$0.2301.0×100%–
prism-ml/ternary-bonsai-2-27bprism-mlopen· 262K ctx3$0.500kilo$0.5001.0×33%–
novita/ernie-4.5-vl-424b-a47bnovita· 123K ctx6$1.25openrouter$1.251.0×0%–
openai/gpt-astra-latestopenai· 1.1M ctx3$50.00kilo$50.001.0×100%yes
aion-labs/aion-2.0aion-labs· 131K ctx3$1.60nano-gpt$1.601.0×100%–
baidu/ernie-4.5-300b-a47bbaidu· 131K ctx2$1.10siliconflow$1.101.0×0%–
cognitivecomputations/dolphin-mistral-24b-venice-editioncognitivecomputations· 128K ctx2$0.900kilo$0.9001.0×0%–
meta/llama-guard-4-12bmetaopen· 128K ctx2$0.180openrouter$0.1801.0×0%–
xai/grok-4-20-beta-0309-non-reasoningxai· 2M ctx2$6.00llmgateway-providers$6.001.0×100%yes
anthropic/claude-opus-4.7-thinkinganthropic· 1M ctx2$25.00302ai$25.001.0×50%–
modelsdev/claude37sonnetmodelsdev· 200K ctx2$15.00helicone$15.001.0×100%–
modelsdev/codestral2501modelsdev· 256K ctx2$0.900azure$0.9001.0×0%–
inclusionai/ring-2.6-1tinclusionai· 262K ctx3$2.50requesty$2.501.0×67%–
modelsdev/glm52shortmodelsdevopen· 200K ctx2$4.50neuralwatt$4.501.0×100%–
zai/glm-4.5-xzai· 128K ctx2$8.90llmgateway-providers$8.901.0×100%–
perplexity/sonar-pro-searchperplexity· 200K ctx2$15.00kilo$15.001.0×0%–
modelsdev/claudeopus45@20251101modelsdev· 200K ctx2$25.00google-vertex$25.001.0×100%–
anthropic/claude-3-haikuanthropic· 200K ctx2$1.25openrouter$1.251.0×100%–
openai/gpt-5.6-terra-proopenai· 1.1M ctx3$12.00nano-gpt$12.001.0×67%yes
modelsdev/anthropicclaude3opusmodelsdev· 200K ctx2$75.00digitalocean$75.001.0×100%–
qwen/qwen3-reranker-4bqwenopen· 41K ctx2$0.120evroc$0.1201.0×0%–
xai/grok-3-minixai· 131K ctx2$0.500poe$0.5001.0×100%–
google/gemini-3.5-flash-thinkinggoogle· 1.0M ctx2$9.00302ai$9.001.0×50%–
modelsdev/claudehaiku45@20251001modelsdev· 200K ctx2$5.00google-vertex$5.001.0×100%–
xai/grok-3xai· 131K ctx2$15.00poe$15.001.0×100%–
tencent/hunyuan-a13b-instructtencent· 131K ctx4$0.570siliconflow-cn$0.5701.0×0%–
sao10k/l3-lunaris-8bsao10k· 8K ctx2$0.050openrouter$0.0501.0×0%–
zai/glm-4.5-airxzai· 128K ctx2$4.50llmgateway$4.501.0×100%–
modelsdev/gemma412bitmodelsdevopen· 262K ctx2$0.250nano-gpt$0.2501.0×100%–
qwen/qwen3-vl-8b-thinkingqwen· 131K ctx2$2.10openrouter$2.101.0×0%–
openai/gpt-6-astra-proopenai· 1.1M ctx3$50.00openrouter$50.001.0×100%yes
inclusionai/ling-2.6-flashinclusionaiopen· 262K ctx2$0.300novita-ai$0.3001.0×50%–
mistralai/mistral-medium-3.1mistralai· 131K ctx3$2.00nano-gpt$2.001.0×100%–
modelsdev/mistralpixtrallarge25020modelsdev· 128K ctx2$6.00edenai$6.001.0×0%–
qwen/qwen3-vl-32b-thinkingqwen· 262K ctx2$1.50siliconflow-cn$1.501.0×0%–
relace/relace-searchrelace· 256K ctx2$3.00openrouter$3.001.0×0%–
openai/gpt-sol-latestopenai· 1.1M ctx3$10.00openrouter$10.001.0×100%yes
sakana/fugu-maxsakana· 1M ctx6$6.00openrouter$6.001.0×100%–
aion-labs/aion-3.0aion-labs· 131K ctx3$6.00kilo$6.001.0×100%–
Jevtypesafe· 64K ctx5—cloudflare-ai-gateway——0%–
Gemini Embedding 001google· 8K ctx3—google-vertex——0%–
modelsdev/cohereembedv40modelsdevopen· 128K ctx2—azure-cognitive-services——0%–
modelsdev/textembedding3largemodelsdev· 8K ctx4—openai——0%–
modelsdev/modelroutermodelsdev· 200K ctx2—azure-cognitive-services——0%–
qwen/qwen3-embedding-4bqwenopen· 41K ctx3—privatemode-ai——0%–
modelsdev/bgemultilingualgemma2modelsdev· 8K ctx2—scaleway——0%–
modelsdev/cohereembedenglishmodelsdevopen· 512 ctx2—azure-cognitive-services——0%–
modelsdev/textembedding3smallmodelsdev· 8K ctx4—openai——0%–
modelsdev/textembeddingada002modelsdev· 8K ctx3—openai——0%–
modelsdev/cohereembedmultilingualmodelsdevopen· 512 ctx2—azure——0%–
Spread is computed on published output $/M for the latest daily snapshot, base tier only. models.dev records no serving precision, so a cheaper provider may be fp8-degraded — every cross-provider view carries a quant: unknown caveat.

Data: models.dev (MIT). Prices are vendor-published; verify before purchase. Methodology.