Best Reasoning & Thinking LLM APIs Compared
Reasoning models allocate extra test-time compute tokens to solve hard coding, math, and planning tasks. Compare their input/output rates carefully—high output ratios matter most here.
| # | Model | Provider | Input / 1M | Output / 1M | Blended | Context |
|---|---|---|---|---|---|---|
| 1 |
inclusionAI: Ling 3.2 Flash (free)
FreeReasoningVision
|
Inclusionai | Free | Free | Free | 1.05M |
| 2 |
Qwen: Qwen3.7 Flash
ReasoningVisionOpen Weight
|
Qwen | $0.03 | $0.13 | $0.055 | 1M |
| 3 |
IBM: Granite 4.2 8B
Reasoning
|
Ibm Granite | $0.06 | $0.15 | $0.083 | 131K |
| 4 |
Meta: Muse Spark 1.2 Contributor
ReasoningVision
|
Meta | $0.10 | $0.20 | $0.125 | 1.05M |
| 5 |
Meta: Muse Spark 1.3 Contributor
ReasoningVision
|
Meta | $0.10 | $0.20 | $0.125 | 1.05M |
| 6 |
OpenAI: gpt-oss-safeguard-20b
Reasoning
|
OpenAI | $0.075 | $0.30 | $0.131 | 131K |
| 7 |
Google: Gemini 2.5 Flash Lite
ReasoningVision
|
$0.10 | $0.40 | $0.175 | 1.05M | |
| 8 |
PrismML: Ternary Bonsai 2 27B
ReasoningVision
|
Prism Ml | $0.075 | $0.50 | $0.181 | 262K |
| 9 |
LiquidAI: LFM2.5-24B-A2B
Reasoning
|
Liquid | $0.05 | $0.70 | $0.212 | 128K |
| 10 |
Qwen: Qwen3.8 Flash
ReasoningVisionOpen Weight
|
Qwen | $0.15 | $0.47 | $0.23 | 1M |
| 11 |
Qwen: Qwen3.8 Omni Flash
ReasoningVisionOpen Weight
|
Qwen | $0.15 | $0.47 | $0.23 | 1M |
| 12 |
KT: Mi:dm K 3.0 Pro
Reasoning
|
Kt | $0.20 | $0.80 | $0.35 | 262K |
| 13 |
Z.ai: GLM 5.3
Reasoning
|
Z Ai | $0.07 | $1.40 | $0.402 | 1.05M |
| 14 |
DeepSeek: DeepSeek V3.1
ReasoningOpen Weight
|
DeepSeek | $0.25 | $0.95 | $0.425 | 164K |
| 15 |
Reflection: Beam 1.0 501B A23B
Reasoning
|
Reflection | $0.25 | $1.00 | $0.438 | 1.05M |
| 16 |
Thinking Machines: Inkling 2 Small
ReasoningVision
|
Thinkingmachines | $0.30 | $1.00 | $0.475 | 1.05M |
| 17 |
Perceptron: Perceptron Mk1.5
ReasoningVision
|
Perceptron | $0.15 | $1.50 | $0.487 | 37K |
| 18 |
RadixArk: DaisyRose 27B
ReasoningVision
|
Radixark | $0.15 | $1.88 | $0.581 | 262K |
| 19 |
Amazon: Nova 2 Lite
ReasoningVision
|
Amazon | $0.30 | $2.50 | $0.85 | 1M |
| 20 |
DeepSeek: R1 0528
ReasoningOpen Weight
|
DeepSeek | $0.571 | $2.29 | $1.00 | 164K |
| 21 |
Z.ai: GLM 5.2
Reasoning
|
Z Ai | $0.62 | $2.42 | $1.07 | 1.05M |
| 22 |
MoonshotAI: Kimi K2 Thinking
Reasoning
|
Moonshotai | $0.60 | $2.50 | $1.07 | 262K |
| 23 |
Elorian: Glimpse
ReasoningVision
|
Elorian | $0.60 | $3.00 | $1.20 | 262K |
| 24 |
Qwen: Qwen3 235B A22B Thinking 2507
ReasoningOpen Weight
|
Qwen | $0.45 | $3.50 | $1.21 | 128K |
| 25 |
Nous: Hermes 4 405B
ReasoningOpen Weight
|
Nous Research | $1.00 | $3.00 | $1.50 | 131K |
| 26 |
SpaceXSI: Grok 4.20
ReasoningVision
|
xAI | $1.25 | $2.50 | $1.56 | 2M |
| 27 |
SpaceXSI: Grok 4.3
ReasoningVision
|
xAI | $1.25 | $2.50 | $1.56 | 1M |
| 28 |
Sakana: Sakana Namazu
ReasoningVision
|
Sakana | $0.95 | $4.00 | $1.71 | 262K |
| 29 |
Meta: Muse Spark 1.1
ReasoningVision
|
Meta | $1.25 | $4.25 | $2.00 | 1.05M |
| 30 |
Meta: Muse Spark 1.2
ReasoningVision
|
Meta | $1.25 | $4.25 | $2.00 | 1.05M |
| 31 |
Meta: Muse Spark 1.3
ReasoningVision
|
Meta | $1.25 | $4.25 | $2.00 | 1.05M |
| 32 |
MoonshotAI: Kimi K3.1 Pro
ReasoningVision
|
Moonshotai | $0.253 | $8.80 | $2.39 | 1.05M |
| 33 |
OpenAI: o3
ReasoningVision
|
OpenAI | $2.00 | $8.00 | $3.50 | 200K |
| 34 |
Perplexity: Sonar Reasoning Pro
ReasoningVision
|
Perplexity | $2.00 | $8.00 | $3.50 | 128K |
| 35 |
Google: Gemini 4 Argon
ReasoningVision
|
$2.00 | $10.00 | $4.00 | 1.05M | |
| 36 |
MoonshotAI: Kimi K3
ReasoningVision
|
Moonshotai | $0.192 | $16.00 | $4.14 | 1.05M |
| 37 |
Google: Gemini 3.1 Pro Preview
ReasoningVision
|
$2.00 | $12.00 | $4.50 | 1.05M | |
| 38 |
Fireworks: Ember-1
ReasoningVision
|
Fireworks | $3.00 | $15.00 | $6.00 | 1.05M |
| 39 |
Meta: Muse Wonder
ReasoningVision
|
Meta | $4.00 | $20.00 | $8.00 | 1.05M |
| 40 |
Anthropic: Claude Opus 4.5
ReasoningVision
|
Anthropic | $5.00 | $25.00 | $10.00 | 200K |
| 41 |
OpenAI: o3 Pro
ReasoningVision
|
OpenAI | $20.00 | $80.00 | $35.00 | 200K |
Try These Models Without Multiple Billing Accounts
OpenRouter
Unified API gateway for 300+ LLMs with automatic fallbacks and unified billing.
Together AI
High-throughput serverless inference & fine-tuning for Llama, DeepSeek, and Qwen.
GroqCloud
LPU-accelerated inference delivering hundreds of tokens per second for real-time voice & agents.
Some outbound infrastructure and API gateway links on KostoZero are partner referrals that help keep this observatory free and independent at no extra cost to you.
Get Instant Price-Drop (>10%), New AI Bot & SaaS Cliff Alerts
Every 6 hours our autonomous sync checks 40+ LLMs, AI crawlers, and SaaS pricing cliffs—alerting you when prices drop >10% or new scrapers appear.