Benchmarks / Cost

Blended Price (USD per 1M tokens, 3:1)

List price per 1M tokens blended 3:1 input:output as tracked by AA. A proxy for cost, not a measured run cost.

251models
–Open Index weight
–last updated

Blended Price (USD per 1M tokens, 3:1)

Lower is better · best effort per model

Chart settings
251 of 251 selected

full table below

#ModelEffortScoreCost / taskHarness
1Gemma 4 E4B (Reasoning)–0.04 USD––
2Gemma 4 E4B (Non-reasoning)none0.04 USD––
3Llama 3.1 8bnone0.04 USD––
4Sarvam 30B (High)high0.05 USD––
5Granite 4.2 3B–0.05 USD––
6Mistral Small 3none0.06 USD––
7Qwen 3.5 4B–0.06 USD––
8Nova Micronone0.06 USD––
9NVIDIA Nemotron Nano 9B V2 (Reasoning)–0.07 USD––
10Llama 3 8bnone0.07 USD––
11Sarvam 105B (High)high0.07 USD––
12Granite 3.3 8B (Non-reasoning)none0.08 USD––
13NVIDIA Nemotron Nano 9B V2 (Non-reasoning)none0.09 USD––
14NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)–0.09 USD––
15NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)none0.09 USD––
16Qwen 2.5 Turbonone0.09 USD––
17Mistral Small 3.1none0.09 USD––
18Nemotron 3.5 Lightning–0.10 USD––
19gpt-oss-20bhigh0.10 USD––
20Hy3–0.10 USD––
21Llama 2 7bnone0.10 USD––
22Ministral 3 3Bnone0.10 USD––
23Nova Litenone0.10 USD––
24Mistral Small 3.2none0.11 USD––
25Granite 4.2 8B–0.11 USD––
26Granite 4.0 H Smallnone0.11 USD––
27Ling 3.0 Flash Fin–0.11 USD––
28Ling 3.0 Flash–0.11 USD––
29DeepSeek V4 Flash 0420none0.12 USD––
30Olmo 3 7B Instructnone0.13 USD––
31Apertus 8B Instructnone0.13 USD––
32Phi-4 Mini Instructnone0.13 USD––
33GPT-5 Nanohigh0.14 USD––
34Phi-4 Multimodal Instructnone0.14 USD––
35Qwen 3 14Bnone0.15 USD––
36Solar Mininone0.15 USD––
37Ministral 3 8Bnone0.15 USD––
38Agnes 2.5 Pro Beta–0.15 USD––
39MiMo V2 Flash–0.15 USD––
40Step 3.5 Flash 2603–0.15 USD––
41Step 3.5 Flash–0.15 USD––
42Gemma 4 12B–0.15 USD––
43Qwen 3.5 9B–0.15 USD––
44GLM 4.7 Flash–0.15 USD––
45Mistralnone0.16 USD––
46MiMo V2.6 Flash–0.18 USD––
47Solar Mini 4–0.18 USD––
48MiMo V2.5–0.18 USD––
49Gemma 4 26B–0.18 USD––
50Gemini 2.5 Flash Lite–0.18 USD––
51GPT-4.1 Nanonone0.18 USD––
52Haiku 5.5max0.20 USD––
53GPT-6 Lunamax0.20 USD––
54Ministral 3 14Bnone0.20 USD––
55Qwen 3 8Bnone0.20 USD––
56Qwen 3 30B-A3Bnone0.21 USD––
57Phi-4none0.22 USD––
58Gemma 4 31Bnone0.22 USD––
59Qwen 3.8 Flash–0.23 USD––
60GLM 5.3 Flashmax0.24 USD––
61gpt-oss-120blow0.24 USD––
62Ling-flash-2.0none0.25 USD––
63Ring-flash-2.0–0.25 USD––
64Jamba 1.5 Mininone0.25 USD––
65Mistral Small 4–0.26 USD––
66Solar Pro 3high0.26 USD––
67Qwen 3 32Bnone0.26 USD––
68GPT-4o mininone0.26 USD––
69Qwen 3.5none0.28 USD––
70Grok 4 Fast–0.28 USD––
71Granite 4.2 30B–0.28 USD––
72Nemotron 3.5 Super VL–0.29 USD––
73Qwen 3 Codernone0.30 USD––
74Seed-OSS-36B-Instruct–0.30 USD––
75Llama 4 Scoutnone0.30 USD––
76NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)none0.30 USD––
77DeepSeek V3.2–0.32 USD––
78Celeris-1none0.33 USD––
79Llama 3.2 11bnone0.34 USD––
80Grok 3 Minihigh0.35 USD––
81Reka Flash (Sep '24)none0.35 USD––
82Reka Flash 3–0.35 USD––
83GLM 4.5 Air–0.37 USD––
84Mercury 2.5–0.38 USD––
85Mercury 2high0.38 USD––
86Qwen 3 235B-A22Bnone0.39 USD––
87Qwen 3none0.39 USD––
88Trinity Large Thinking–0.39 USD––
89DeepSeek V3none0.41 USD––
90Llama 4 Mavericknone0.41 USD––
91Step 3.7 Flashhigh0.44 USD––
92Ling 3.1 Flash–0.45 USD––
93Nemotron 3 Super 120B A12B (Reasoning)–0.45 USD––
94GLM 4.6 V–0.45 USD––
95Nemotron 3 Nano Omni 30B A3B Reasoning–0.45 USD––
96GPT-5.6 Lunamax0.45 USD––
97DeepSeek V3.1none0.45 USD––
98GPT-5.4 Nanoxhigh0.46 USD––
99Qwen 2.5 72Bnone0.48 USD––
100ERNIE 4.5 300B A47Bnone0.49 USD––
101Haiku 3none0.50 USD––
102Mixtral 8x7B Instructnone0.51 USD––
103DeepSeek V4.1 Flashmax0.52 USD––
104MiniMax M3–0.52 USD––
105Solar Pro 4–0.52 USD––
106Inkling Small–0.52 USD––
107MiniMax M2.5–0.52 USD––
108MiniMax M2.7–0.52 USD––
109KAT Coder Pro V2none0.52 USD––
110MiniMax M2.1–0.52 USD––
111MiniMax M2–0.52 USD––
112Muse Glimmer (High)high0.52 USD––
113MiMo V2.6 Pro–0.54 USD––
114DeepSeek V4 Pro 0424max0.54 USD––
115MiMo V2.5 Pro–0.54 USD––
116Llama 3.1 70bnone0.56 USD––
117Gemini 3.1 Flash-Litehigh0.56 USD––
118DeepSeek V4 Flashmax0.66 USD––
119DeepSeek V4 Flash 0731max0.66 USD––
120GPT-5 Minimedium0.69 USD––
121GPT-5.1 Codex Minihigh0.69 USD––
122Qwen 3.5 35B-A3B–0.69 USD––
123Qwen 3.7 Plus–0.70 USD––
124GPT-4.1 Mininone0.70 USD––
125Llama 3.3 70bnone0.71 USD––
126Mistral Large 3none0.75 USD––
127Magistral Small 1.2–0.75 USD––
128GPT-3.5none0.75 USD––
129DeepSeek R1–0.80 USD––
130Mistral Medium 3none0.80 USD––
131Qwen 3.5 27B–0.82 USD––
132Qwen 3.6 35B-A3B–0.84 USD––
133Gemini 3.5 Flash-Litehigh0.85 USD––
134Ling-2.6-1Tnone0.85 USD––
135Ring-2.6-1Txhigh0.85 USD––
136Nova 2.0 Omni (Medium)medium0.85 USD––
137Nova 2.0 Lite (High)high0.85 USD––
138Gemini 2.5 Flash–0.85 USD––
139Nova 2.0 Lite (Medium)medium0.85 USD––
140Nova 2.0 Lite (Low)low0.85 USD––
141Nova 2.0 Omni (Low)low0.85 USD––
142Nova 2.0 Lite (Non-reasoning)none0.85 USD––
143Nova 2.0 Omni (Non-reasoning)none0.85 USD––
144Quasar 438B (Max, Based on GLM-5.2)max0.90 USD––
145GLM 4.5 V–0.90 USD––
146GLM 4.6–0.96 USD––
147MiniMax M1–0.96 USD––
148Apodex 1.1–0.97 USD––
149GLM 4.7–1.00 USD––
150Kimi K2none1.00 USD––
151Nemotron 3 Ultramax1.07 USD––
152Qwen 3.5 122B-A10Bnone1.10 USD––
153Qwen 3.8 27Bxhigh1.13 USD––
154Qwen 3.6 Plus–1.13 USD––
155Gemini 3 Flash–1.13 USD––
156Kimi K2.5–1.14 USD––
157Llama 3 70bnone1.18 USD––
158Grok Build 0.1–1.25 USD––
159Cogito v2.1 (Reasoning)–1.25 USD––
160LongCat 2.0–1.30 USD––
161Apertus 70B Instructnone1.34 USD––
162Qwen 3.5 397B-A17Bnone1.35 USD––
163Qwen 3.6 27B–1.35 USD––
164Nova Pronone1.40 USD––
165Step 5 Preview–1.43 USD––
166Gemini 3.8 Flashhigh1.50 USD––
167Gemini 3.7 Flashmedium1.50 USD––
168Gemini 3.6 Flashhigh1.50 USD––
169Llama 3.1 405b–1.50 USD––
170GLM 5–1.55 USD––
171Grok 4.20–1.56 USD––
172Grok 4.3high1.56 USD––
173GPT-5.4 Minixhigh1.69 USD––
174Kimi K2.6–1.71 USD––
175Kimi K2.7 Code–1.71 USD––
176Inklingxhigh1.76 USD––
177o4 minihigh1.93 USD––
178o3 mini–1.93 USD––
179DeepSeek V4 Pro 0813max1.98 USD––
180GLM 5.1–1.98 USD––
181Muse Spark 1.3max2.00 USD––
182Muse Spark 1.2xhigh2.00 USD––
183Muse Spark Open (Xhigh)xhigh2.00 USD––
184Muse Spark 1.1xhigh2.00 USD––
185Haiku 4.5–2.00 USD––
186Mistral Large 4–2.06 USD––
187GLM 5.3max2.15 USD––
188GLM 5.2max2.15 USD––
189Grok 4.7xhigh3.00 USD––
190Qwen 3.8 Maxmax3.00 USD––
191Grok 4.6high3.00 USD––
192Qwen 3.8 2.4T-A95B–3.00 USD––
193Grok 4.5high3.00 USD––
194Mistral Medium 3.5high3.00 USD––
195Mistral Large 2none3.00 USD––
196Gemini 3.5 Flashmedium3.38 USD––
197GPT-5 Codexhigh3.44 USD––
198GPT-5.1high3.44 USD––
199GPT-5.1 Codexhigh3.44 USD––
200GPT-5high3.44 USD––
201Gemini 2.5 Pro–3.44 USD––
202Nova 2.0 Pro Preview (Medium)medium3.44 USD––
203Nova 2.0 Pro Preview (Low)low3.44 USD––
204Nova 2.0 Pro Preview (Non-reasoning)none3.44 USD––
205o3–3.50 USD––
206GPT-4.1none3.50 USD––
207Jamba 1.5 Largenone3.50 USD––
208Qwen 3.7 Maxmax3.75 USD––
209Sonnet 5.5max4.00 USD––
210Gemini 4high4.00 USD––
211GPT-6.1 Solmax4.00 USD––
212GPT-6 Solmax4.00 USD––
213Sonnet 5max4.00 USD––
214GPT-4onone4.38 USD––
215Command Anone4.38 USD––
216GPT-5.6 Terramax4.50 USD––
217Gemini 3.1 Pro–4.50 USD––
218Gemini 3 Prohigh4.50 USD––
219GPT-5.3 Codexxhigh4.81 USD––
220GPT-5.2xhigh4.81 USD––
221GPT-5.2 Codexxhigh4.81 USD––
222GPT-5.4xhigh5.63 USD––
223Kimi K3max6.00 USD––
224Sonnet 4.6max6.00 USD––
225Grok 4–6.00 USD––
226Sonnet 4.5–6.00 USD––
227Sonnet 3.7none6.00 USD––
228Grok 3none6.00 USD––
229Sonnet 3.5none6.00 USD––
230Mistral Largenone6.00 USD––
231Opus 5.5max8.00 USD––
232Muse Wondermax8.00 USD––
233GPT-5.6 Solmax8.00 USD––
234Opus 5max10.00 USD––
235Opus 4.8max10.00 USD––
236Opus 4.7max10.00 USD––
237Opus 4.6max10.00 USD––
238Opus 4.5–10.00 USD––
239GPT-5.5xhigh11.25 USD––
240GPT-5.5 Instant–11.25 USD––
241GPT-4none15.00 USD––
242Fable 5.1max20.00 USD––
243GPT-6 Astramax20.00 USD––
244Fable 5max20.00 USD––
245o1–26.25 USD––
246Opus 4.1–30.00 USD––
247Opus 4–30.00 USD––
248Opus 3none30.00 USD––
249o3 pro–35.00 USD––
250GPT-5.4 Proxhigh67.50 USD––
251o1 pro–262.50 USD––