Best LLM APIs for Reasoning Tasks
LLM APIs specialized for reasoning and complex problem solving. Models ranked by GPQA Diamond score, with verified pricing for math, logic, and multi-step inference.
MODELPRICEWATCH.COM · 2026-08-09
Cost calculator for this use casemonthly cost, top 3 models
- 1 Claude Sonnet 5
- $—
- 2 GPT-5.6 Sol
- $—
- 3 Gemini 3.1 Pro
- $—
Full ranking — top 25 models
list prices, USD per 1M tokens| Model | GPQA Diamond | Blended* | Input | Output | Context | Provider |
|---|---|---|---|---|---|---|
| 1Claude Sonnet 5 | 96.2% | $4.00 | $2.00 | $10.00 | 1M | Anthropic |
| 2GPT-5.6 Sol | 94.6% | $11.25 | $5.00 | $30.00 | 1M | OpenAI |
| 3Gemini 3.1 Pro | 94.3% | $4.50 | $2.00 | $12.00 | 2M | |
| 4Claude Opus 4.7 | 94.2% | $10.00 | $5.00 | $25.00 | 1M | Anthropic |
| 5Claude Fable 5 | 94.1% | $20.00 | $10.00 | $50.00 | 1M | Anthropic |
| 6Claude Opus 4.8 | 93.6% | $10.00 | $5.00 | $25.00 | 1M | Anthropic |
| 7GPT-5.5 | 93.6% | $11.25 | $5.00 | $30.00 | 1M | OpenAI |
| 8Kimi K3 | 93.5% | $6.00 | $3.00 | $15.00 | 1M | Moonshot |
| 9MiniMax-M3 | 93% | $0.525 | $0.300 | $1.20 | 1M | MiniMax |
| 10GPT-5.6 Terra | 92.9% | $4.50 | $2.00 | $12.00 | 1M | OpenAI |
| 11GPT-5.2 | 92.4% | $4.81 | $1.75 | $14.00 | 400K | OpenAI |
| 12GPT-5.6 Luna | 92.3% | $0.450 | $0.200 | $1.20 | 1M | OpenAI |
| 13Claude Opus 4.6 | 91.3% | $10.00 | $5.00 | $25.00 | 1M | Anthropic |
| 14GLM-5.2 | 91.2% | $2.15 | $1.40 | $4.40 | 1M | Z.AI |
| 15Kimi K2.6 | 90.5% | $1.71 | $0.950 | $4.00 | 262K | Moonshot |
| 16DeepSeek V4 Pro | 90.1% | $0.544 | $0.435 | $0.870 | 1M | DeepSeek |
| 17Claude Sonnet 4.6 | 89.9% | $6.00 | $3.00 | $15.00 | 1M | Anthropic |
| 18DeepSeek V4 Flash | 88.1% | $0.175 | $0.140 | $0.280 | 1M | DeepSeek |
| 19GPT-5.4 | 88% | $5.63 | $2.50 | $15.00 | 1M | OpenAI |
| 20Kimi K2.5 | 87.6% | $1.20 | $0.600 | $3.00 | 262K | Moonshot |
| 21Qwen3.8-Max | 92.6% vendor | $3.00 | $2.00 | $6.00 | 1M | Alibaba |
| 22Inkling Small | — | $0.525 | $0.300 | $1.20 | 262K | Thinking Machines |
| 23Inkling | — | $1.76 | $1.00 | $4.05 | 262K | Thinking Machines |
| 24Muse Spark 1.2 | — | $2.00 | $1.25 | $4.25 | 1M | Meta |
| 25Claude Opus 5 | — | $10.00 | $5.00 | $25.00 | 1M | Anthropic |
Recent price movement in this ranking
price-only deltas · logged by the daily scan3 of the top 25 reasoning models have re-priced since we began tracking · last checked Aug 9, 2026.
| Model | Changes | Latest move | Provider |
|---|---|---|---|
| GPT-5.6 Terra | 1 | price cut on Aug 1, 2026: $2.5/$15 → $2/$12 /Mtok | OpenAI |
| GPT-5.6 Luna | 1 | price cut on Aug 1, 2026: $1/$6 → $0.2/$1.2 /Mtok | OpenAI |
| MiniMax-M3 | 1 | price cut on Jun 17, 2026: $0.6/$2.4 → $0.3/$1.2 /Mtok | MiniMax |
Most recently, GPT-5.6 Terra cut its price on Aug 1, 2026 — a sign pricing in this category is still moving, so re-check before committing to a long-term choice. See all recent price moves →
How models are selected
Generally-available models with a GPQA Diamond score (graduate-level science questions) or a reasoning-focused release, ranked by independently-evaluated GPQA Diamond. One row per model — host duplicates are collapsed. Vendor-reported scores are shown flagged and ranked below independently-scored models, never against them; models with no published score rank last, cheapest first.
Prices are per million tokens (Mtok), sourced directly from — and linked to — official provider pricing pages. "Blended cost" is (3×input + 1×output) ÷ 4 — weighted toward input because real workloads read far more tokens than they generate.