Skip to main content

Best Value LLM for Reasoning in 2026 — Cost-Adjusted Rankings

Reasoning models tend to be the most expensive tier — they use chain-of-thought, produce more output tokens, and are priced accordingly. This ranking divides each model's weighted reasoning score by output token price, revealing which models deliver the best abstract reasoning, long-context comprehension, and multi-step logic per dollar. For applications that need strong reasoning without frontier-model budgets, the value leaders here are worth serious consideration.

Unless noted otherwise, ranking surfaces on this page use BenchLM's provisional leaderboard lane rather than the stricter sourced-only verified leaderboard.

Bottom line: Reasoning models are expensive — chain-of-thought generates more output tokens. GPT-4.1 nano and Gemini 3.1 Flash-Lite offer the best reasoning per dollar.

According to BenchLM.ai, GPT-4.1 nano leads this ranking with a score of 166.72, followed by Gemini 3.1 Flash-Lite (149.7) and Gemini 2.5 Flash (74.78). There is a significant gap between the leading models and the rest of the field.

The best open-weight option is DeepSeek Coder 2.0 (ranked #4 with a score of 54.71). While proprietary models lead, open-weight options are within striking distance for teams willing to trade a few points of performance for full model control.

This ranking is based on provisional weighted averages across the scoring benchmarks in reasoning tracked by BenchLM.ai. For detailed model profiles, click any model name below. To compare two specific models head-to-head, use the "vs #" links.

What changed

GPT-4.1 nano leads reasoning value — best reasoning capability per dollar.

Gemini 3.1 Flash-Lite close second on reasoning value at the lowest price point.

Gemini 2.5 Flash good reasoning value with broader capabilities.

How to choose

Full Rankings (30 models)

GPT-4.1 nano
OpenAI·Proprietary·1M

166.72

Score/$

Score: 66.7 · $0.4/1M

Gemini 3.1 Flash-Lite
Google·Proprietary·1M

149.7

Score/$

Score: 59.9 · $0.4/1M

Gemini 2.5 Flash
Google·Proprietary·1M

74.78

Score/$

Score: 44.9 · $0.6/1M

4
DeepSeek Coder 2.0
DeepSeek·Open Weight·128K

54.71

Score/$

Score: 60.2 · $1.1/1M

5
GPT-4.1 mini
OpenAI·Proprietary·1M

49.61

Score/$

Score: 79.4 · $1.6/1M

6
GPT-4o mini
OpenAI·Proprietary·128K

40.12

Score/$

Score: 24.1 · $0.6/1M

7
Gemini 3 Flash
Google·Proprietary·1M

23.01

Score/$

Score: 69 · $3/1M

8
Kimi K2.5
Moonshot AI·Open Weight·128K

19.8

Score/$

Score: 55.4 · $2.8/1M

9
Gemini 3.1 Pro
Google·Proprietary·1M

19.39

Score/$

Score: 97 · $5/1M

10
o3-mini
OpenAI·Proprietary·200K

15.49

Score/$

Score: 68.2 · $4.4/1M

11
Claude Haiku 4.5
Anthropic·Proprietary·200K

15.02

Score/$

Score: 60.1 · $4/1M

12
GLM-5.1
Z.AI·Open Weight·203K

14.73

Score/$

Score: 64.8 · $4.4/1M

13
Gemini 2.5 Pro
Google·Proprietary·1M

11.95

Score/$

Score: 59.7 · $5/1M

14
GPT-5.1
OpenAI·Proprietary·200K

11.34

Score/$

Score: 68 · $6/1M

15
GPT-5.1-Codex-Max
OpenAI·Proprietary·400K

11.24

Score/$

Score: 89.9 · $8/1M

16
GPT-5.2-Codex
OpenAI·Proprietary·400K

11.07

Score/$

Score: 88.5 · $8/1M

17
DeepSeek-R1
DeepSeek·Open Weight·128K

10.85

Score/$

Score: 23.8 · $2.19/1M

18
GPT-5.2
OpenAI·Proprietary·400K

10.69

Score/$

Score: 85.5 · $8/1M

19
GPT-4.1
OpenAI·Proprietary·1M

9.69

Score/$

Score: 77.5 · $8/1M

20
GPT-5.3 Codex
OpenAI·Proprietary·400K

9.47

Score/$

Score: 94.7 · $10/1M

21
Mistral Large 3
Mistral·Proprietary·128K

8.96

Score/$

Score: 53.8 · $6/1M

22
GPT-5.4
OpenAI·Proprietary·1.05M

6.2

Score/$

Score: 93 · $15/1M

23
Grok 4.1
xAI·Proprietary·1M

6.13

Score/$

Score: 91.9 · $15/1M

24
Claude Sonnet 4.6
Anthropic·Proprietary·200K

5.52

Score/$

Score: 82.8 · $15/1M

25
GPT-4o
OpenAI·Proprietary·128K

4.63

Score/$

Score: 46.3 · $10/1M

26
Claude Sonnet 4.5
Anthropic·Proprietary·200K

4.2

Score/$

Score: 63 · $15/1M

27
o3
OpenAI·Proprietary·200K

1.4

Score/$

Score: 55.9 · $40/1M

28
o1
OpenAI·Proprietary·200K

1.24

Score/$

Score: 74.2 · $60/1M

29
Claude Opus 4.6
Anthropic·Proprietary·1M

1.2

Score/$

Score: 90 · $75/1M

30
o1-pro
OpenAI·Proprietary·200K

0.06

Score/$

Score: 33.1 · $600/1M

These rankings update weekly

Get notified when models move. One email a week with what changed and why.

Free. No spam. Unsubscribe anytime.

Key Takeaways

The best value model is GPT-4.1 nano by OpenAI with a provisional Score/$ ratio of 166.72 (score: 66.7, output: $0.4/1M tokens).

The best open-weight model is DeepSeek Coder 2.0 at position #4.

30 models are included in this ranking.

Score in Context

What these scores mean

Value scores divide the weighted reasoning score by output token price (per 1M tokens). Higher means more capability per dollar. Models with no listed price are excluded.

Known limitations

Value rankings favor cheap models even if absolute performance is modest. A model scoring half as well at one-tenth the price wins on value — but may not meet your quality bar. Always check raw scores alongside value rankings.

Last updated: April 10, 2026

The AI models change fast. We track them for you.

For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.

Free. No spam. Unsubscribe anytime.