Skip to main content
BenchLM
Advertise

Cost intelligence

LLM API Pricing Comparison 2026

Compare API pricing for every major LLM side by side. All prices are per million tokens. The Score/$column uses the public overall score per dollar of output — higher means better value.

Last updated: July 23, 2026. Prices reflect official API rates. Open-weight model costs depend on your infrastructure. How does LLM token pricing work?

The cheapest LLM API right now is Ministral 3 3B at $0.1/$0.1 per million input/output tokens. The cheapest production-grade model (BenchLM score 70+) is Grok 4.5 at $2/$6. The cheapest frontier-tier model (80+) is Kimi K3 at $3/$15.

Provider pricing hubs:Claude API pricingOpenAI API pricingGemini API pricingDeepSeek API pricingGrok API pricingKimi API pricingQwen API pricingMistral API pricing
Cheapest API

Ministral 3 3B

$0.1 in / $0.1 out per 1M tokens

Best Value (Score/$)

Ministral 3 3B (Reasoning)

Score/$ ratio: 395.2

Highest Scored (Paid)

Claude Mythos 5

Public score: 83.93 — $10 in / $50 out

Model the full break-even →
CSVJSON
Model / providerInput · output
Claude Mythos 5Anthropic · 1M+Score 83.93 · 1.7 score/$
$10.00·$50.00
Claude Fable 5Anthropic · 1M+Score 83.68 · 1.7 score/$
$10.00·$50.00
GPT-5.6 SolOpenAI · 1MScore 81.96 · 2.7 score/$
$5.00·$30.00
Kimi K3Moonshot AI · 1.05MScore 80.96 · 5.4 score/$
$3.00·$15.00
Claude Opus 4.8Anthropic · 1MScore 78.34 · 3.1 score/$
$5.00·$25.00
Grok 4.5xAI · 500KScore 76.72 · 12.8 score/$
$2.00·$6.00
GPT-5.4OpenAI · 1.05MScore 74.24 · 5.0 score/$
$2.50·$15.00
GPT-5.5OpenAI · 1MScore ~73.51 · 2.5 score/$
$5.00·$30.00
Qwen3.7 MaxAlibaba · 1MScore 72.84
Not listed
GPT-5.6 TerraOpenAI · 1MScore ~72.57 · 4.8 score/$
$2.50·$15.00
Claude Opus 4.7Anthropic · 1MScore 71.94 · 2.9 score/$
$5.00·$25.00
MiMo-V2.5-ProXiaomi · 1MScore 70.19
Not listed
MiniMax M3MiniMax · 1MScore 69.75 · 58.1 score/$
$0.30·$1.20
Claude Opus 4.6Anthropic · 1MScore 68.59 · 2.7 score/$
$5.00·$25.00
GLM-5.1Z.AI · 203KScore 67.74 · 15.4 score/$
$1.40·$4.40
Gemini 3 ProGoogle · 2MScore 67.73 · 5.6 score/$
$2.00·$12.00
InklingThinking Machines Lab · 1MScore 67.54 · 14.4 score/$
$1.87·$4.68
GPT-5.6 LunaOpenAI · 1MScore ~67.17 · 11.2 score/$
$1.00·$6.00
GPT-5.2 ProOpenAI · 400KScore 67.01 · 0.5 score/$
$25.00·$150.00
GLM-5-TurboZ.AI · 200KScore 66.89 · 16.7 score/$
$1.20·$4.00
GPT-5.4 nanoOpenAI · 400KScore 66.79 · 53.4 score/$
$0.20·$1.25
GPT-5.3 CodexOpenAI · 400KScore 66.69 · 4.8 score/$
$1.75·$14.00
Claude Opus 4.7 (Adaptive)Anthropic · 1MScore ~66.27 · 2.6 score/$
$5.00·$25.00
GLM-5Z.AI · 200KScore 66.06 · 20.6 score/$
$1.00·$3.20
Claude Sonnet 5Anthropic · 1MScore ~65.32 · 6.5 score/$
$2.00·$10.00
Qwen3.6 PlusAlibaba · 1MScore 65.2
Not listed
Grok 4.3xAI · 1MScore 65.1 · 26.0 score/$
$1.25·$2.50
Claude Sonnet 4.6Anthropic · 200KScore 65.07 · 4.3 score/$
$3.00·$15.00
Gemini 3.5 FlashGoogle · 1MScore ~64.75 · 7.2 score/$
$1.50·$9.00
Claude Opus 4.5Anthropic · 200KScore 64.22 · 2.6 score/$
$5.00·$25.00
MiniMax M2.7MiniMax · 200KScore 64.11 · 53.4 score/$
$0.30·$1.20
GLM-5.2Z.AI · 1MScore ~63.96 · 14.5 score/$
$1.40·$4.40
GPT-5.5 ProOpenAI · 1MScore ~63.69 · 0.3 score/$
$30.00·$180.00
GLM-5V-TurboZ.AI · 200KScore 63.5 · 15.9 score/$
$1.20·$4.00
Gemini 3 Pro Deep ThinkGoogle · 2MScore ~61.31
Not listed
GLM-4.7Z.AI · 200KScore 61.16
Free*
Gemma 4 31BGoogle · 256KScore 61.08
Free*
GPT-5.4 ProOpenAI · 1.05MScore ~60.89 · 0.3 score/$
$30.00·$180.00
Qwen3.5-27BAlibaba · 262KScore 60.7
Free*
DeepSeek V4 ProDeepSeek · 1MScore 60.66 · 69.7 score/$
$0.43·$0.87
Qwen3.5-122B-A10BAlibaba · 262KScore 60.56
Free*
Gemini 3 FlashGoogle · 1MScore 60.49 · 20.2 score/$
$0.50·$3.00
Grok 4xAI · 128KScore 60.42
Not listed
Grok 4.1xAI · 1MScore ~59.97
Not listed
GLM-5 (Reasoning)Z.AI · 200KScore ~59.77 · 18.7 score/$
$1.00·$3.20
Kimi K2.5Moonshot AI · 256KScore 59.66 · 19.9 score/$
$0.60·$3.00
MiniMax M2.5MiniMax · 128KScore 59.52 · 49.6 score/$
$0.30·$1.20
Qwen3.5 397B (Reasoning)Alibaba · 128KScore ~59.5 · 16.5 score/$
$0.60·$3.60
Kimi K2.5 (Reasoning)Moonshot AI · 256KScore ~59.35 · 19.8 score/$
$0.60·$3.00
GPT-5.2-CodexOpenAI · 400KScore 59.1 · 4.2 score/$
$1.75·$14.00
GPT-5.2 InstantOpenAI · 128KScore ~59 · 9.8 score/$
$1.50·$6.00
GPT-5.3 InstantOpenAI · 128KScore ~58.9 · 4.2 score/$
$1.75·$14.00
DeepSeek V4 FlashDeepSeek · 1MScore ~58.88 · 210.3 score/$
$0.14·$0.28
MiMo-V2.5Xiaomi · 1MScore ~58.62
Not listed
GPT-5 (high)OpenAI · 400KScore ~58.61 · 5.9 score/$
$1.25·$10.00
GPT-5.2OpenAI · 400KScore ~58.43 · 4.2 score/$
$1.75·$14.00
DeepSeek V3.2 (Thinking)DeepSeek · 128KScore ~58.15 · 26.6 score/$
$0.55·$2.19
Qwen3 235B 2507 (Reasoning)Alibaba · 128KScore ~58.01
Free*
Gemma 4 26B A4BGoogle · 256KScore 57.96
Free*
GLM-4.5Z.AI · 128KScore ~57.56 · 26.2 score/$
$0.60·$2.20
Gemini 2.5 ProGoogle · 1MScore 57.25 · 5.7 score/$
$1.25·$10.00
Qwen3.5 397BAlibaba · 128KScore ~57.01 · 15.8 score/$
$0.60·$3.60
Qwen3.5-35B-A3BAlibaba · 262KScore 56.97
Free*
GPT-5.3-Codex-SparkOpenAI · 256KScore ~56.91
Not listed
Kimi 2.6Moonshot AI · 256KScore ~56.79 · 14.2 score/$
$0.95·$4.00
GPT-5.4 miniOpenAI · 400KScore ~56.77 · 12.6 score/$
$0.75·$4.50
Claude Haiku 4.5Anthropic · 200KScore ~56.58 · 11.3 score/$
$1.00·$5.00
Qwen3 235B 2507Alibaba · 128KScore ~56.02
Free*
Trinity-Large-PreviewArcee AI · 512KScore ~55.94 · 55.9 score/$
$0.25·$1.00
Hy3Tencent · 256KScore ~55.64
Free*
DeepSeek V4 Pro (High)DeepSeek · 1MScore ~55.47 · 63.8 score/$
$0.43·$0.87
DeepSeek V3.2DeepSeek · 128KScore 55.4 · 131.9 score/$
$0.28·$0.42
Gemini 3.1 ProGoogle · 1MScore ~55.3 · 4.6 score/$
$2.00·$12.00
GPT-5 (medium)OpenAI · 128KScore 55.15
Not listed
Step 3.5 FlashStepFun · 256KScore 55.1 · 183.7 score/$
$0.10·$0.30
Kimi K2.7 CodeMoonshot AI · 256KScore ~55 · 13.8 score/$
$0.95·$4.00
Grok 4.20xAI · 2MScore ~54.68 · 9.1 score/$
$2.00·$6.00
DeepSeek LLM 2.0DeepSeek · 128KScore ~54.53
Free*
GPT-5.1-Codex-MaxOpenAI · 400KScore ~54.48 · 5.5 score/$
$1.25·$10.00
MiMo-V2-FlashXiaomi · 256KScore 54.06
Free*
DeepSeek V4 Flash (High)DeepSeek · 1MScore ~53.95 · 192.7 score/$
$0.14·$0.28
Qwen3.6-27BAlibaba · 262KScore ~53.82
Free*
GPT-5.1OpenAI · 400KScore ~53.65 · 5.4 score/$
$1.25·$10.00
DeepSeek V3.1DeepSeek · 128KScore 53.64
Free*
Claude Sonnet 4.5Anthropic · 200KScore ~53.61 · 3.6 score/$
$3.00·$15.00
DeepSeek V3.1 (Reasoning)DeepSeek · 128KScore 53.43
Free*
Nemotron 3 Nano 30BNVIDIA · 32KScore ~52.94
Free*
Trinity-Large-ThinkingArcee AI · 512KScore 52.32 · 58.1 score/$
$0.25·$0.90
Qwen2.5-72BAlibaba · 128KScore ~52.15
Free*
Llama 3.1 405BMeta · 128KScore ~51.71
Free*
DeepSeek R1DeepSeek · 128KScore 51.67 · 23.6 score/$
$0.55·$2.19
Mercury 2Inception · 128KScore 51.28 · 68.4 score/$
$0.25·$0.75
GLM-4.7-FlashZ.AI · 200KScore 51.25
Free*
Grok 4.1 FastxAI · 2MScore 51.25 · 102.5 score/$
$0.20·$0.50
GPT-4.1OpenAI · 1MScore 51.11 · 6.4 score/$
$2.00·$8.00
Nemotron 3 Super 120B A12BNVIDIA · 256KScore ~50.97
Free*
Step 3.7 FlashStepFun · 256KScore ~50.87 · 44.2 score/$
$0.20·$1.15
Gemini 3.1 Flash-LiteGoogle · 1MScore 50.83 · 33.9 score/$
$0.25·$1.50
Llama 3 70BMeta · 128KScore ~50.82
Free*
Mistral Large 3Mistral · 256KScore 50.4 · 33.6 score/$
$0.50·$1.50
DeepSeek Coder 2.0DeepSeek · 128KScore ~50.28
Not listed
Seed 1.6ByteDance · 256KScore ~50.23
Not listed
GPT-OSS 120BOpenAI · 128KScore 50.08
Free*
Nemotron 3 Super 100BNVIDIA · 1MScore ~50.08
Free*
o4-mini (high)OpenAI · 200KScore ~49.98
Not listed
DeepSeekMath V2DeepSeek · 128KScore ~49.89
Free*
Qwen2.5-1MAlibaba · 1MScore ~49.89
Free*
Seed-2.0-LiteByteDance · 256KScore ~49.84
Not listed
Ministral 3 14B (Reasoning)Mistral · 256KScore ~49.34 · 246.7 score/$
$0.20·$0.20
o1-previewOpenAI · 200KScore 49.11 · 0.8 score/$
$15.00·$60.00
Aion-2.0Aion Labs · 128KScore ~48.66 · 30.4 score/$
$0.80·$1.60
Mixtral 8x22B Instruct v0.1Mistral · 64KScore ~48.54
Free*
o3-proOpenAI · 200KScore 48.33 · 0.6 score/$
$20.00·$80.00
o1OpenAI · 200KScore ~48.1 · 0.8 score/$
$15.00·$60.00
Gemini 2.5 FlashGoogle · 1MScore 48.09 · 19.2 score/$
$0.30·$2.50
o3OpenAI · 200KScore 47.89 · 6.0 score/$
$2.00·$8.00
Claude 3.5 SonnetAnthropic · 200KScore ~47.74 · 3.2 score/$
$3.00·$15.00
GLM-4.5-AirZ.AI · 128KScore 47.7 · 43.4 score/$
$0.20·$1.10
Qwen3.5 FlashAlibaba · 1MScore 47.65 · 119.1 score/$
$0.10·$0.40
Command A+Cohere · 128KScore ~47.51 · 4.8 score/$
$2.50·$10.00
o3-miniOpenAI · 200KScore 47.41 · 10.8 score/$
$1.10·$4.40
Qwen3.5 PlusAlibaba · 1MScore ~47.2 · 19.7 score/$
$0.40·$2.40
GPT-5 nanoOpenAI · 400KScore ~46.36 · 115.9 score/$
$0.05·$0.40
Mistral Small 4Mistral · 256KScore ~46.23 · 77.0 score/$
$0.15·$0.60
o1-proOpenAI · 200KScore ~45.94 · 0.1 score/$
$150.00·$600.00
Claude 4.1 OpusAnthropic · 200KScore 45.88 · 0.6 score/$
$15.00·$75.00
Z-1Z · 128KScore ~45.13
Not listed
Nemotron-4 15BNVIDIA · 32KScore ~45.08
Free*
Seed 1.6 FlashByteDance · 256KScore ~45.08
Not listed
Mistral 8x7BMistral · 32KScore ~44.99
Free*
DeepSeek V3DeepSeek · 128KScore 44.97 · 40.9 score/$
$0.27·$1.10
Moonshot v1Moonshot AI · 128KScore ~44.79
Not listed
Seed-2.0-MiniByteDance · 256KScore ~44.6
Not listed
Nemotron Ultra 253BNVIDIA · 32KScore ~44.4
Free*
Nemotron 3 Nano Omni 30B A3BNVIDIA · 256KScore ~44.24
Free*
GPT-4.1 miniOpenAI · 1MScore ~44.19 · 27.6 score/$
$0.40·$1.60
GPT-5 miniOpenAI · 128KScore 43.93 · 22.0 score/$
$0.25·$2.00
Ling 2.6 FlashInclusionAI · 262KScore ~43.87
Not listed
Gemma 4 E4BGoogle · 128KScore ~43.2
Free*
Mistral Medium 3Mistral · 128KScore ~43.2 · 21.6 score/$
$0.40·$2.00
Sarvam 105BSarvam · 128KScore ~42.97
Free*
Claude 4 SonnetAnthropic · 200KScore 42.79 · 2.9 score/$
$3.00·$15.00
GPT-OSS 20BOpenAI · 128KScore 42.74
Free*
DeepSeek R1 Distill Qwen 32BDeepSeek · 128KScore ~42.58
Free*
GPT-4.1 nanoOpenAI · 1MScore ~42.06 · 105.2 score/$
$0.10·$0.40
Gemma 4 E2BGoogle · 128KScore ~41.82
Free*
Mistral Large 2Mistral · 128KScore ~41.77
Not listed
Gemma 3 27BGoogle · 32KScore 41.57
Free*
GPT-4oOpenAI · 128KScore 41.49 · 4.2 score/$
$2.50·$10.00
LFM2.5-8B-A1BLiquidAI · 128KScore ~41.42
Free*
Claude 3 OpusAnthropic · 200KScore 41.13 · 0.6 score/$
$15.00·$75.00
Sarvam 30BSarvam · 64KScore ~40.7
Free*
Grok 3 [Beta]xAI · 128KScore 40.43
Not listed
Ministral 3 8B (Reasoning)Mistral · 256KScore ~40.38 · 269.2 score/$
$0.15·$0.15
Mistral 7B v0.3Mistral · 32KScore ~39.9
Free*
Llama 4 ScoutMeta · 10MScore 39.87
Free*
Qwen2.5-VL-32BAlibaba · 32KScore ~39.85
Free*
Llama 4 BehemothMeta · 32KScore ~39.8
Free*
Ministral 3 3B (Reasoning)Mistral · 256KScore ~39.52 · 395.2 score/$
$0.10·$0.10
Mistral 8x7B v0.2Mistral · 32KScore ~39.13
Free*
Grok Code Fast 1xAI · 256KScore 38.64 · 25.8 score/$
$0.20·$1.50
Granite-4.0-350MIBM · 32KScore ~38.32
Free*
Granite-4.0-H-350MIBM · 32KScore ~38.32
Free*
GPT-4o miniOpenAI · 128KScore 37.87 · 63.1 score/$
$0.15·$0.60
Claude 4.1 Opus ThinkingAnthropic · 200KScore 36.55
Not listed
Gemini 1.5 ProGoogle · 1MScore 35.71 · 7.1 score/$
$1.25·$5.00
Qwen2.5 Coder 32B InstructAlibaba · 128KScore 34.7
Free*
Ministral 3 14BMistral · 256KScore 34.47 · 172.3 score/$
$0.20·$0.20
GPT-4 TurboOpenAI · 128KScore 27.44 · 0.9 score/$
$10.00·$30.00
Kimi K2Moonshot AI · 128KScore 27.19 · 10.9 score/$
$0.60·$2.50
MiniMax M1 80kMiniMax · 80KScore 25.12
Not listed
Llama 4 MaverickMeta · 1MScore 23.49
Free*
Phi-4Microsoft · 16KScore 22.69
Free*
Gemini 1.0 ProGoogle · 32KScore 21.78
Not listed
Claude 3 HaikuAnthropic · 200KScore 21.39 · 17.1 score/$
$0.25·$1.25
Ministral 3 8BMistral · 256KScore 20.96 · 139.7 score/$
$0.15·$0.15
Nova ProAmazon · 128KScore 20.32
Not listed
LFM2-24B-A2BLiquidAI · 32KScore 18.87
Free*
Ministral 3 3BMistral · 256KScore 18.27 · 182.7 score/$
$0.10·$0.10
LFM2.5-1.2B-ThinkingLiquidAI · 32KScore 16.23
Free*
LFM2.5-1.2B-InstructLiquidAI · 32KScore 15.52
Free*
1-bit Bonsai 1.7BPrism ML · 32KScore not ranked
Free*
1-bit Bonsai 4BPrism ML · 32KScore not ranked
Free*
1-bit Bonsai 8BPrism ML · 64KScore not ranked
Free*
Composer 2Cursor · 200KScore not ranked
$0.50·$2.50
Composer 2.5Cursor · 200KScore not ranked
$0.50·$2.50
Cosmos3-EdgeNVIDIA · 256KScore not ranked
Free*
DBRX InstructDatabricks · 32KScore not ranked
Free*
DeepSeek V4 Flash (Max)DeepSeek · 1MScore not ranked
$0.14·$0.28
DeepSeek V4 Flash BaseDeepSeek · 1MScore not ranked
Not listed
DeepSeek V4 Pro (Max)DeepSeek · 1MScore not ranked
$0.43·$0.87
DeepSeek V4 Pro BaseDeepSeek · 1MScore not ranked
Not listed
Fugu CyberSakana AI · 1MScore not ranked
$6.00·$36.00
Gemini 3.5 Flash-LiteGoogle · 1MScore not ranked
$0.30·$2.50
Gemini 3.6 FlashGoogle · 1MScore not ranked
$1.50·$7.50
Granite-4.0-1BIBM · 128KScore not ranked
Free*
Granite-4.0-H-1BIBM · 128KScore not ranked
Free*
Grok 3 MinixAI · 128KScore not ranked
$0.30·$0.50
Grok 4.20 Multi-agentxAI · 2MScore not ranked
Not listed
Grok Build 0.1xAI · 256KScore not ranked
$1.00·$2.00
Holo3-122B-A10BH Company · 64KScore not ranked
$0.40·$3.00
Holo3-35B-A3BH Company · 64KScore not ranked
Not listed
Holo3.1-0.8BH Company · 262KScore not ranked
Free*
Holo3.1-35B-A3BH Company · 64KScore not ranked
$0.25·$1.80
Holo3.1-35B-A3B-FP8H Company · 262KScore not ranked
Free*
Holo3.1-35B-A3B-GGUFH Company · 262KScore not ranked
Free*
Holo3.1-35B-A3B-NVFP4H Company · 262KScore not ranked
Free*
Holo3.1-4BH Company · 262KScore not ranked
Free*
Holo3.1-9BH Company · 262KScore not ranked
Free*
Hy3 PreviewTencent · 256KScore not ranked
Free*
Interfaze BetaInterfaze · 1MScore not ranked
$1.50·$3.50
Laguna M.1Poolside · 256KScore not ranked
Not listed
Laguna S 2.1Poolside · 1MScore not ranked
$0.10·$0.20
Laguna XS.2Poolside · 256KScore not ranked
Free*
LeanstralMistral · 256KScore not ranked
Free*
LFM2.5-230MLiquidAI · 32KScore not ranked
Free*
LFM2.5-350MLiquidAI · 32KScore not ranked
Free*
LFM2.5-ColBERT-350MLiquidAI · 32KScore not ranked
Free*
LFM2.5-Embedding-350MLiquidAI · 32KScore not ranked
Free*
LFM2.5-VL-450MLiquidAI · 128KScore not ranked
Free*
LongCat-2.0Meituan · 1MScore not ranked
Free*
Mistral Medium 3.5 128BMistral · 256KScore not ranked
$1.50·$7.50
Mistral Small 4 (Reasoning)Mistral · 256KScore not ranked
$0.15·$0.60
Nemotron 3 UltraNVIDIA · 1MScore not ranked
Free*
o4-miniOpenAI · 200KScore not ranked
$1.10·$4.40
Ornith-1.0-35BDeepReinforce AI · 256KScore not ranked
Free*
Ornith-1.0-397BDeepReinforce AI · 256KScore not ranked
Free*
Ornith-1.0-9BDeepReinforce AI · 256KScore not ranked
Free*
Qwen3.8 Max PreviewAlibaba · Context not listedScore not ranked
Not listed
Sakana FuguSakana AI · 1MScore not ranked
Not listed
Sakana Fugu-UltraSakana AI · 1MScore not ranked
Not listed
Soofi S 30B-A3BSoofi Project · 1MScore not ranked
Free*
Ternary Bonsai 1.7BPrism ML · 32KScore not ranked
Free*
Ternary Bonsai 4BPrism ML · 32KScore not ranked
Free*
Ternary Bonsai 8BPrism ML · 64KScore not ranked
Free*
ZAYA1-74B-PreviewZyphra · 256KScore not ranked
Free*
ZAYA1-8BZyphra · 131KScore not ranked
Free*

* Score/$ = Overall benchmark score ÷ output price per million tokens. Higher is better. Free/open-weight models are excluded.

Price vs Performance

Interactive scatter chart with efficiency frontier

Cost Calculator

Estimate cost per blog post, page, or feature

Token Counter

Count tokens for any text across models

Best Value by Category

Best Value: Coding→Best Value: Agentic→Best Value: Reasoning→Best Value: Knowledge→Best Value: Math→Best Value: Multimodal→Best Value: Overall→

Frequently Asked Questions

Which LLM has the best price-to-performance ratio?

Based on our benchmarks, models like DeepSeek V3, Gemini 3 Flash, and Gemini 3.1 Flash-Lite offer strong price-to-performance ratios. DeepSeek V3 scores competitively on benchmarks while costing a fraction of proprietary frontier models, and Google's Flash tiers are positioned for lower-cost high-volume usage. For free self-hosted options, Meta's Llama 4 models provide excellent performance at zero API cost.

How much more expensive is GPT-5.4 Pro than GPT-5.4?

GPT-5.4 is $2.50 input and $15.00 output per million tokens. GPT-5.4 Pro is $30.00 input and $180.00 output. That makes Pro 12x the cost on output tokens, so it only makes sense if the accuracy bump is worth a very real increase in spend.

Are open-source LLMs really free?

Open-weight models like Llama 4, DeepSeek, and Qwen are free to download, but running them requires GPU infrastructure. Self-hosting costs vary from $0.50-$5.00/hour for cloud GPUs depending on model size. Many providers also offer API access to open models at lower prices than proprietary alternatives.

What is the cheapest LLM API in 2026?

Gemini 3.1 Flash-Lite is among the cheapest proprietary APIs at $0.10/$0.40 per million tokens (input/output). Gemini 2.5 Flash remains very affordable at $0.15/$0.60, Gemini 3 Flash currently sits at $0.50/$3.00, and DeepSeek V3 is also inexpensive at $0.27/$1.10. For reasoning models, o3-mini and o4-mini offer budget-friendly options at $1.10/$4.40.

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.