Benchmark suiteLatest source data: Oct 16, 2026Checked: October 18, 2026

Coding model rankings.

A focused coding view using Code Arena and SWE-bench today, with clear source slots for Terminal-Bench and Aider-style coding benchmarks as those feeds are wired.

What These Mean

Ranked models

Code Arena or SWE-bench signal

SWE-bench matches

Matched to model rows

Top coding index

Claude Opus 5.5

85 coding index

Best SWE-bench

Pending

Coding formula

Code generation plus real issue resolution.

The coding index blends Code Arena and SWE-bench where available, with output speed as a small tie-breaker. Terminal-Bench and Aider remain visible as planned feed slots until we have a reliable ingestion path.

Coding performance

Higher coding index is better. SWE-bench cells show verified model matches only.

Top 30 of 90

Rank	Model	Index	Code Arena	SWE-bench	Speed	Run mode	Sources
#1	Claude Opus 5.5 Claude Opus 5.5 (High) AnthropicProprietary	85coding index	1,807Code Arena rank #1	Not listedNo confident match	78 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#2	GPT-6 Astra GPT-6 Astra (Max) OpenAIProprietary	78coding index	1,777Code Arena rank #2	Not listedNo confident match	57 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#3	GPT-6.1 Sol GPT-6.1 Sol (Max) OpenAIProprietary	73coding index	1,755Code Arena rank #4	Not listedNo confident match	56 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#4	GPT-6 Sol GPT-6 Sol (Max) OpenAIProprietary	71coding index	1,683Code Arena rank #8	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#5	Claude Fable 5.1 Claude Fable 5.1 (Max) AnthropicProprietary	71coding index	1,741Code Arena rank #5	Not listedNo confident match	56 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#6	Gemini 4 Argon Gemini 4 Argon (High) GoogleProprietary	69coding index	1,678Code Arena rank #9	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#7	Qwen3.8 Max qwen3.8-max AlibabaProprietary	68coding index	1,672Code Arena rank #11	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#8	Claude Sonnet 5.5 Claude Sonnet 5.5 (xHigh) AnthropicProprietary	67coding index	1,711Code Arena rank #6	Not listedNo confident match	100 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#9	Claude Opus 5 Claude Opus 5 (High) AnthropicProprietary	64coding index	1,657Code Arena rank #13	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#10	DeepSeek V4.1 Flash DeepSeek-V4.1-Flash (Max) DeepSeekOpen weights	60coding index	1,618Code Arena rank #24	Not listedNo confident match	337 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#11	Muse Wonder Muse Wonder (Max) MetaProprietary	58coding index	1,650Code Arena rank #15	Not listedNo confident match	173 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#12	Hy4 Preview Hy4 preview TencentOpen weights	58coding index	1,631Code Arena rank #18	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-bench
#13	Muse Spark 1.3 Muse Spark 1.3 (Max) MetaProprietary	57coding index	1,622Code Arena rank #22	Not listedNo confident match	266 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#14	Claude Fable 5 Claude Fable 5 (High) AnthropicProprietary	57coding index	1,625Code Arena rank #21	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#15	Grok 4.6 Grok 4.6 (High) xAIProprietary	55coding index	1,617Code Arena rank #25	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#16	GPT-5.6 Sol GPT-5.6 Sol (xHigh, Codex Harness) OpenAIProprietary	55coding index	1,616Code Arena rank #26	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#17	Kimi K3 Kimi K3 (Max) Moonshot AIProprietary	53coding index	1,655Code Arena rank #14	Not listedNo confident match	40 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#18	GLM 5.2 GLM-5.2 (Max) Z.aiOpen weights	52coding index	1,605Code Arena rank #28	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#19	Grok 4.7 Grok 4.7 (xHigh) xAIProprietary	51coding index	1,639Code Arena rank #16	Not listedNo confident match	73 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#20	MiMo V2.6 Flash mimo-v2.6-flash XiaomiOpen weights	49coding index	1,634Code Arena rank #17	Not listedNo confident match	50 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#21	Gemini 3.7 Flash Gemini 3.7 Flash (High) GoogleProprietary	49coding index	1,591Code Arena rank #30	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#22	Qwen3.8 Flash Next Qwen3.8-Flash-Next AlibabaProprietary	49coding index	1,630Code Arena rank #19	Not listedNo confident match	53 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#23	GLM 5.3 GLM-5.3 (Max) Z.aiOpen weights	48coding index	1,621Code Arena rank #23	Not listedNo confident match	85 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#24	Gemini 3.8 Flash Gemini 3.8 Flash (High) GoogleProprietary	47coding index	1,585Code Arena rank #31	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#25	MiMo V2.6 Pro mimo-v2.6-pro XiaomiOpen weights	47coding index	1,625Code Arena rank #20	Not listedNo confident match	45 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#26	DeepSeek V4 Flash DeepSeek-V4-Flash (High) DeepSeekOpen weights	46coding index	1,581Code Arena rank #34	Not listedNo confident match	Not listedoutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#27	Claude Haiku 5.5 Claude Haiku 5.5 (High) AnthropicProprietary	45coding index	1,585Code Arena rank #32	Not listedNo confident match	169 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#28	GLM 5.3 Flash glm-5.3-flash Z.aiOpen weights	45coding index	1,605Code Arena rank #27	Not listedNo confident match	71 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#29	GPT-6 Luna GPT-6 Luna (Max) OpenAIProprietary	42coding index	1,578Code Arena rank #35	Not listedNo confident match	124 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis
#30	Qwen3.8 27b qwen3.8-27b AlibabaOpen weights	41coding index	1,593Code Arena rank #29	Not listedNo confident match	42 tok/soutput speed	PlannedTerminal-Bench + Aider feed	Code ArenaSWE-benchArtificial Analysis

Claude Opus 5.5

Claude Opus 5.5 (High)

AnthropicProprietary

85coding index

Code Arena

1,807Code Arena rank #1

SWE-bench

Not listedNo confident match

Speed

78 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

GPT-6 Astra

GPT-6 Astra (Max)

OpenAIProprietary

78coding index

Code Arena

1,777Code Arena rank #2

SWE-bench

Not listedNo confident match

Speed

57 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

GPT-6.1 Sol

GPT-6.1 Sol (Max)

OpenAIProprietary

73coding index

Code Arena

1,755Code Arena rank #4

SWE-bench

Not listedNo confident match

Speed

56 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

GPT-6 Sol

GPT-6 Sol (Max)

OpenAIProprietary

71coding index

Code Arena

1,683Code Arena rank #8

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

Claude Fable 5.1

Claude Fable 5.1 (Max)

AnthropicProprietary

71coding index

Code Arena

1,741Code Arena rank #5

SWE-bench

Not listedNo confident match

Speed

56 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

Gemini 4 Argon

Gemini 4 Argon (High)

GoogleProprietary

69coding index

Code Arena

1,678Code Arena rank #9

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

Qwen3.8 Max

qwen3.8-max

AlibabaProprietary

68coding index

Code Arena

1,672Code Arena rank #11

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

Claude Sonnet 5.5

Claude Sonnet 5.5 (xHigh)

AnthropicProprietary

67coding index

Code Arena

1,711Code Arena rank #6

SWE-bench

Not listedNo confident match

Speed

100 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

Claude Opus 5

Claude Opus 5 (High)

AnthropicProprietary

64coding index

Code Arena

1,657Code Arena rank #13

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#10

DeepSeek V4.1 Flash

DeepSeek-V4.1-Flash (Max)

DeepSeekOpen weights

60coding index

Code Arena

1,618Code Arena rank #24

SWE-bench

Not listedNo confident match

Speed

337 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#11

Muse Wonder

Muse Wonder (Max)

MetaProprietary

58coding index

Code Arena

1,650Code Arena rank #15

SWE-bench

Not listedNo confident match

Speed

173 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#12

Hy4 Preview

Hy4 preview

TencentOpen weights

58coding index

Code Arena

1,631Code Arena rank #18

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-bench

#13

Muse Spark 1.3

Muse Spark 1.3 (Max)

MetaProprietary

57coding index

Code Arena

1,622Code Arena rank #22

SWE-bench

Not listedNo confident match

Speed

266 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#14

Claude Fable 5

Claude Fable 5 (High)

AnthropicProprietary

57coding index

Code Arena

1,625Code Arena rank #21

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#15

Grok 4.6

Grok 4.6 (High)

xAIProprietary

55coding index

Code Arena

1,617Code Arena rank #25

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#16

GPT-5.6 Sol

GPT-5.6 Sol (xHigh, Codex Harness)

OpenAIProprietary

55coding index

Code Arena

1,616Code Arena rank #26

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#17

Kimi K3

Kimi K3 (Max)

Moonshot AIProprietary

53coding index

Code Arena

1,655Code Arena rank #14

SWE-bench

Not listedNo confident match

Speed

40 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#18

GLM 5.2

GLM-5.2 (Max)

Z.aiOpen weights

52coding index

Code Arena

1,605Code Arena rank #28

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#19

Grok 4.7

Grok 4.7 (xHigh)

xAIProprietary

51coding index

Code Arena

1,639Code Arena rank #16

SWE-bench

Not listedNo confident match

Speed

73 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#20

MiMo V2.6 Flash

mimo-v2.6-flash

XiaomiOpen weights

49coding index

Code Arena

1,634Code Arena rank #17

SWE-bench

Not listedNo confident match

Speed

50 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#21

Gemini 3.7 Flash

Gemini 3.7 Flash (High)

GoogleProprietary

49coding index

Code Arena

1,591Code Arena rank #30

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#22

Qwen3.8 Flash Next

Qwen3.8-Flash-Next

AlibabaProprietary

49coding index

Code Arena

1,630Code Arena rank #19

SWE-bench

Not listedNo confident match

Speed

53 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#23

GLM 5.3

GLM-5.3 (Max)

Z.aiOpen weights

48coding index

Code Arena

1,621Code Arena rank #23

SWE-bench

Not listedNo confident match

Speed

85 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#24

Gemini 3.8 Flash

Gemini 3.8 Flash (High)

GoogleProprietary

47coding index

Code Arena

1,585Code Arena rank #31

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#25

MiMo V2.6 Pro

mimo-v2.6-pro

XiaomiOpen weights

47coding index

Code Arena

1,625Code Arena rank #20

SWE-bench

Not listedNo confident match

Speed

45 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#26

DeepSeek V4 Flash

DeepSeek-V4-Flash (High)

DeepSeekOpen weights

46coding index

Code Arena

1,581Code Arena rank #34

SWE-bench

Not listedNo confident match

Speed

Not listedoutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#27

Claude Haiku 5.5

Claude Haiku 5.5 (High)

AnthropicProprietary

45coding index

Code Arena

1,585Code Arena rank #32

SWE-bench

Not listedNo confident match

Speed

169 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#28

GLM 5.3 Flash

glm-5.3-flash

Z.aiOpen weights

45coding index

Code Arena

1,605Code Arena rank #27

SWE-bench

Not listedNo confident match

Speed

71 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#29

GPT-6 Luna

GPT-6 Luna (Max)

OpenAIProprietary

42coding index

Code Arena

1,578Code Arena rank #35

SWE-bench

Not listedNo confident match

Speed

124 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

#30

Qwen3.8 27b

qwen3.8-27b

AlibabaOpen weights

41coding index

Code Arena

1,593Code Arena rank #29

SWE-bench

Not listedNo confident match

Speed

42 tok/soutput speed

Run mode

PlannedTerminal-Bench + Aider feed

Code ArenaSWE-benchArtificial Analysis

Benchmark guide

What the scores mean.

A quick reading key for comparing coding models without confusing source coverage, preference scores, and real issue-resolution benchmarks.

Higher: coding indexMissing rows are not zero

What does the coding index measure?

The coding index combines Code Arena and SWE-bench signals where available, with output speed used as a small tie-breaker. It is a practical coding comparison, not a guarantee that the model will solve every repository task.

How should I read Code Arena versus SWE-bench?

Code Arena is a public preference-style benchmark for coding outputs, while SWE-bench measures real software issue resolution. Code Arena is broader and more available; SWE-bench is more task-specific and harder to match across model names.

Why does Terminal-Bench show as planned?

Terminal-Bench is relevant for agentic coding workflows, but this page does not mix it into rankings until a reliable source feed is wired. Planned benchmarks are called out so the table does not imply hidden or invented scores.

Why are some SWE-bench cells empty?

Empty SWE-bench cells mean there is no public model row or no confident match to the model name in this table. They should be read as missing source coverage, not as a zero percent result.