Skip to main content
BenchLM
Advertise

BenchLM recommendation

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

Data verified

As of July 23, 2026, the top model in best value llm for coding on the BenchLM leaderboard is Ministral 3 3B with a score of 213.

Last verified: July 23, 2026

Raw benchmark scores only tell half the story. This ranking divides each model's weighted coding score by its output token price, surfacing models that deliver the most coding capability per dollar spent. A model scoring 70 at $1/1M tokens outranks one scoring 80 at $15/1M tokens here — because for the same budget you get far more coding work done. Use this alongside the standard coding leaderboard to find the sweet spot between performance and cost for your coding workflows.

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

Bottom line: Gemini 3.1 Flash-Lite dominates coding value at $0.40/1M output. DeepSeek Coder 2.0 offers the best absolute coding performance per dollar among serious coding models.

Ministral 3 3B leads this ranking with a score of 213, followed by DeepSeek V4 Flash (Max) (185.75) and DeepSeek V4 Flash (High) (185.04). There is a significant gap between the leading models and the rest of the field.

The best open-weight option is Ministral 3 3B (ranked #1 with a score of 213). Open-weight models are highly competitive in this category — self-hosting is a viable alternative to proprietary APIs.

This ranking is based on provisional weighted averages across the scoring benchmarks in coding 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.

1Open

Ministral 3 3B

Mistral · 128K

213Score/$

Score: 21.3 · $0.1/1M

2Open

DeepSeek V4 Flash (Max)

DeepSeek · 1M

185.75Score/$

Score: 52 · $0.28/1M

3Open

DeepSeek V4 Flash (High)

DeepSeek · 1M

185.04Score/$

Score: 51.8 · $0.28/1M

What changed

Gemini 3.1 Flash-Lite leads coding value — most coding capability per dollar spent.

DeepSeek Coder 2.0 best value among dedicated coding models with strong raw scores.

Gemini 2.5 Flash good coding value with broader model capabilities.

How to choose

Maximum coding per dollar?

Gemini 3.1 Flash-Lite — lowest cost with decent coding

Strong coding at good value?

DeepSeek Coder 2.0 — best raw coding scores per dollar

Best raw coding regardless of cost?

See the standard coding leaderboard

Full Rankings (81 models)

1
Ministral 3 3B
Mistral·Open Weight·128K

213

Score/$

Score: 21.3 · $0.1/1M

2
DeepSeek V4 Flash (Max)
DeepSeek·Open Weight·1M

185.75

Score/$

Score: 52 · $0.28/1M

3
DeepSeek V4 Flash (High)
DeepSeek·Open Weight·1M

185.04

Score/$

Score: 51.8 · $0.28/1M

4
Ministral 3 14B
Mistral·Open Weight·128K

169.7

Score/$

Score: 33.9 · $0.2/1M

5
DeepSeek V4 Flash
DeepSeek·Open Weight·1M

168.29

Score/$

Score: 47.1 · $0.28/1M

6
Ministral 3 8B
Mistral·Open Weight·128K

160.33

Score/$

Score: 24.1 · $0.15/1M

7
DeepSeek V3.2
DeepSeek·Open Weight·128K

122.4

Score/$

Score: 51.4 · $0.42/1M

8
Qwen3.5 Flash
Alibaba·Proprietary·1M

108.4

Score/$

Score: 43.4 · $0.4/1M

9
GPT-4.1 nano
OpenAI·Proprietary·1M

105.2

Score/$

Score: 42.1 · $0.4/1M

10
Mistral Small 4
Mistral·Open Weight·256K

77.18

Score/$

Score: 46.3 · $0.6/1M

11
DeepSeek V4 Pro (High)
DeepSeek·Open Weight·1M

61.66

Score/$

Score: 53.6 · $0.87/1M

12
GPT-4o mini
OpenAI·Proprietary·128K

59.33

Score/$

Score: 35.6 · $0.6/1M

13
DeepSeek V4 Pro (Max)
DeepSeek·Open Weight·1M

59.09

Score/$

Score: 51.4 · $0.87/1M

14
Mercury 2
Inception·Proprietary·128K

58.08

Score/$

Score: 43.6 · $0.75/1M

15
Trinity-Large-Thinking
Arcee AI·Open Weight·512K

51.8

Score/$

Score: 46.6 · $0.9/1M

16
DeepSeek V4 Pro
DeepSeek·Open Weight·1M

49.97

Score/$

Score: 43.5 · $0.87/1M

17
MiniMax M2.5
MiniMax·Proprietary·128K

44.67

Score/$

Score: 53.6 · $1.2/1M

18
Step 3.7 Flash
StepFun·Open Weight·256K

43.63

Score/$

Score: 50.2 · $1.15/1M

19
MiniMax M3
MiniMax·Open Weight·1M

41.24

Score/$

Score: 49.5 · $1.2/1M

20
MiniMax M2.7
MiniMax·Open Weight·200K

39.64

Score/$

Score: 47.6 · $1.2/1M

21
DeepSeek V3
DeepSeek·Open Weight·128K

38.75

Score/$

Score: 42.6 · $1.1/1M

22
GPT-5.4 nano
OpenAI·Proprietary·400K

34.88

Score/$

Score: 43.6 · $1.25/1M

23
Gemini 3.1 Flash-Lite
Google·Proprietary·1M

29.44

Score/$

Score: 44.2 · $1.5/1M

24
GPT-4.1 mini
OpenAI·Proprietary·1M

27.79

Score/$

Score: 44.5 · $1.6/1M

25
DeepSeek V3.2 (Thinking)
DeepSeek·Open Weight·128K

23.63

Score/$

Score: 51.8 · $2.19/1M

26
Grok Code Fast 1
xAI·Proprietary·256K

22.97

Score/$

Score: 34.5 · $1.5/1M

27
Grok Build 0.1
xAI·Proprietary·256K

21

Score/$

Score: 42 · $2/1M

28
Composer 2
Cursor·Proprietary·200K

20.83

Score/$

Score: 52.1 · $2.5/1M

29
Mistral Large 3
Mistral·Proprietary·128K

20.65

Score/$

Score: 31 · $1.5/1M

30
GPT-5 mini
OpenAI·Proprietary·128K

20.23

Score/$

Score: 40.5 · $2/1M

31
Kimi K2.5
Moonshot AI·Open Weight·256K

19

Score/$

Score: 57 · $3/1M

32
GLM-5
Z.AI·Open Weight·200K

18.56

Score/$

Score: 59.4 · $3.2/1M

33
Kimi K2.5 (Reasoning)
Moonshot AI·Proprietary·128K

17.9

Score/$

Score: 53.7 · $3/1M

34
Gemini 3 Flash
Google·Proprietary·1M

16.06

Score/$

Score: 48.2 · $3/1M

35
Grok 4.3
xAI·Proprietary·1M

15.74

Score/$

Score: 39.4 · $2.5/1M

36
GLM-5.2
Z.AI·Open Weight·1M

14.79

Score/$

Score: 65.1 · $4.4/1M

37
Qwen3.5 397B
Alibaba·Open Weight·128K

14.69

Score/$

Score: 52.9 · $3.6/1M

38
GLM-5V-Turbo
Z.AI·Proprietary·200K

14.66

Score/$

Score: 58.6 · $4/1M

39
Kimi K2.7 Code
Moonshot AI·Open Weight·256K

13.56

Score/$

Score: 54.2 · $4/1M

40
GLM-5.1
Z.AI·Open Weight·203K

13.07

Score/$

Score: 57.5 · $4.4/1M

41
Kimi K2.6
Moonshot AI·Open Weight·256K

13

Score/$

Score: 52 · $4/1M

42
GPT-5.6 Luna
OpenAI·Proprietary·1M

12.09

Score/$

Score: 72.6 · $6/1M

43
o3-mini
OpenAI·Proprietary·200K

10.98

Score/$

Score: 48.3 · $4.4/1M

44
GPT-5.4 mini
OpenAI·Proprietary·400K

10.22

Score/$

Score: 46 · $4.5/1M

45
Grok 4.5
xAI·Proprietary·500K

9.85

Score/$

Score: 59.1 · $6/1M

46
Inkling
Thinking Machines Lab·Open Weight·1M

9.71

Score/$

Score: 45.4 · $4.68/1M

47
Claude Haiku 4.5
Anthropic·Proprietary·200K

9.6

Score/$

Score: 48 · $5/1M

48
Grok 4.20
xAI·Proprietary·2M

7.84

Score/$

Score: 47 · $6/1M

49
Gemini 1.5 Pro
Google·Proprietary·2M

7.32

Score/$

Score: 36.6 · $5/1M

50
Gemini 3.5 Flash
Google·Proprietary·1M

6.89

Score/$

Score: 62 · $9/1M

51
Claude Sonnet 5
Anthropic·Proprietary·1M

6.84

Score/$

Score: 68.4 · $10/1M

52
GPT-4.1
OpenAI·Proprietary·1M

6.15

Score/$

Score: 49.2 · $8/1M

53
Mistral Medium 3.5 128B
Mistral·Open Weight·256K

6.09

Score/$

Score: 45.7 · $7.5/1M

54
GPT-5.1
OpenAI·Proprietary·200K

5.29

Score/$

Score: 52.9 · $10/1M

55
Gemini 3 Pro
Google·Proprietary·2M

5.23

Score/$

Score: 62.8 · $12/1M

56
Kimi K3
Moonshot AI·Pending·1.05M

5.2

Score/$

Score: 78 · $15/1M

57
GPT-5.3 Codex
OpenAI·Proprietary·400K

4.59

Score/$

Score: 64.2 · $14/1M

58
GPT-5.6 Terra
OpenAI·Proprietary·1M

4.4

Score/$

Score: 66 · $15/1M

59
GPT-5.2-Codex
OpenAI·Proprietary·400K

4.13

Score/$

Score: 57.9 · $14/1M

60
Gemini 3.1 Pro
Google·Proprietary·1M

4.12

Score/$

Score: 49.4 · $12/1M

61
GPT-5.4
OpenAI·Proprietary·1.05M

3.97

Score/$

Score: 59.6 · $15/1M

62
Claude Sonnet 4.6
Anthropic·Proprietary·200K

3.97

Score/$

Score: 59.6 · $15/1M

63
GPT-5.2
OpenAI·Proprietary·400K

3.87

Score/$

Score: 54.2 · $14/1M

64
Gemini 2.5 Pro
Google·Proprietary·1M

3.57

Score/$

Score: 35.7 · $10/1M

65
Command A+
Cohere·Open Weight·128K

3.54

Score/$

Score: 35.4 · $10/1M

66
Claude Sonnet 4.5
Anthropic·Proprietary·200K

3.44

Score/$

Score: 51.6 · $15/1M

67
Claude 4 Sonnet
Anthropic·Proprietary·200K

2.92

Score/$

Score: 43.9 · $15/1M

68
Claude Opus 4.8
Anthropic·Proprietary·1M

2.82

Score/$

Score: 70.6 · $25/1M

69
Claude Opus 4.7
Anthropic·Proprietary·1M

2.74

Score/$

Score: 68.4 · $25/1M

70
GPT-5.6 Sol
OpenAI·Proprietary·1M

2.63

Score/$

Score: 78.9 · $30/1M

71
Claude Opus 4.7 (Adaptive)
Anthropic·Proprietary·1M

2.51

Score/$

Score: 62.8 · $25/1M

72
Claude Opus 4.6
Anthropic·Proprietary·1M

2.5

Score/$

Score: 62.5 · $25/1M

73
Claude Opus 4.5
Anthropic·Proprietary·200K

2.48

Score/$

Score: 61.9 · $25/1M

74
GPT-5.5
OpenAI·Proprietary·1M

2.38

Score/$

Score: 71.4 · $30/1M

75
Claude Mythos 5
Anthropic·Proprietary·1M+

1.62

Score/$

Score: 80.8 · $50/1M

76
Claude Fable 5
Anthropic·Proprietary·1M+

1.61

Score/$

Score: 80.6 · $50/1M

77
GPT-4 Turbo
OpenAI·Proprietary·128K

1.02

Score/$

Score: 30.7 · $30/1M

78
o1
OpenAI·Proprietary·200K

0.82

Score/$

Score: 49.4 · $60/1M

79
o1-preview
OpenAI·Proprietary·200K

0.79

Score/$

Score: 47.6 · $60/1M

80
Claude 4.1 Opus
Anthropic·Proprietary·200K

0.62

Score/$

Score: 46.7 · $75/1M

81
Claude 3 Opus
Anthropic·Proprietary·200K

0.53

Score/$

Score: 39.4 · $75/1M

Key Takeaways

The best value model is Ministral 3 3B by Mistral with a provisional Score/$ ratio of 213 (score: 21.3, output: $0.1/1M tokens).

The best open-weight model is Ministral 3 3B at position #1.

81 models are included in this ranking.

Score in Context

What these scores mean

Value scores divide the weighted coding 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.

Explore More

Last updated: July 23, 2026

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.