Workload-Fit Ranking

50 tracked text models · 65 tracked · 62 in catalog

Rank AI Models for Your Workload

Select a preset or fine-tune weights for coding, reasoning, speed, cost, and context across 50 tracked text models. Image, video and audio models are not ranked here: they have no coding or reasoning rating to weigh. Rankings and suitability scores recompute in real time.

Workload Presets

Choose a preset or adjust fine-tuning sliders below

Fine-tune weightsCode 60 · Reason 70 · Cost 50 · Speed 40 · Ctx 30

Coding rating60%

CompareLLM Coding rating

Deep Reasoning70%

CompareLLM Reasoning rating

Cost Efficiency50%

Low token price per 1M

Output Speed40%

Tokens generated per sec

Context Window30%

Token capacity for large documents

Preference Persistence

Your sliders are automatically saved locally and personalize your future sessions.

Ranked Results (50 Models)

Showing 1–25 of 50·

50 total

GPT-6 AstraBest match
OpenAI
68.8
Fit score
Why this rank?
Coding 24Reasoning 28Cost 0Speed 5Context 12
Reasoning accounts for the largest share of this model's score.
Claude Opus 5
Anthropic
68.6
Fit score
Why this rank?
Coding 24Reasoning 26Cost 2Speed 5Context 12
Reasoning accounts for the largest share of this model's score.
GPT-6 Sol
OpenAI
67.6
Fit score
Why this rank?
Coding 22Reasoning 23Cost 5Speed 6Context 12
Reasoning accounts for the largest share of this model's score.
Claude Fable 5.1
Anthropic
65.1
Fit score
Why this rank?
Coding 22Reasoning 28Cost 0Speed 4Context 12
Reasoning accounts for the largest share of this model's score.
GPT-6 Luna
OpenAI
60.2
Fit score
Why this rank?
Coding 13Reasoning 14Cost 14Speed 7Context 12
Reasoning accounts for the largest share of this model's score.
GPT-5.6 Luna
OpenAI
59.6
Fit score
Why this rank?
Coding 13Reasoning 16Cost 11Speed 7Context 12
Reasoning accounts for the largest share of this model's score.
Claude Opus 5.5
Anthropic
59.2
Fit score
Why this rank?
Coding 24Reasoning 14Cost 3Speed 6Context 12
Coding accounts for the largest share of this model's score.
DeepSeek V4.1 Flash
DeepSeek
59.1
Fit score
Why this rank?
Coding 15Reasoning 8Cost 13Speed 11Context 12
Coding accounts for the largest share of this model's score.
GLM 5.3 Flash
Z.ai
57.6
Fit score
Why this rank?
Coding 16Reasoning 9Cost 16Speed 4Context 12
Value accounts for the largest share of this model's score.
Claude Opus 4.8
Anthropic
57.5
Fit score
Why this rank?
Coding 18Reasoning 19Cost 2Speed 8Context 12
Reasoning accounts for the largest share of this model's score.
GPT-5.6 Terra
OpenAI
57.4
Fit score
Why this rank?
Coding 18Reasoning 19Cost 4Speed 5Context 12
Reasoning accounts for the largest share of this model's score.
Muse Spark 1.3
Meta
57.4
Fit score
Why this rank?
Coding 19Reasoning 11Cost 9Speed 7Context 12
Coding accounts for the largest share of this model's score.
MiMo-V2.6-Pro
Xiaomi
56.6
Fit score
Why this rank?
Coding 17Reasoning 11Cost 14Speed 2Context 12
Coding accounts for the largest share of this model's score.
Gemini 3.7 Flash
Google
55.1
Fit score
Why this rank?
Coding 15Reasoning 8Cost 9Speed 11Context 12
Coding accounts for the largest share of this model's score.
Grok 4.6
xAI
53.9
Fit score
Why this rank?
Coding 21Reasoning 10Cost 8Speed 7Context 8
Coding accounts for the largest share of this model's score.
Grok 4.7
xAI
53.0
Fit score
Why this rank?
Coding 17Reasoning 11Cost 8Speed 9Context 8
Coding accounts for the largest share of this model's score.
Claude Sonnet 5
Anthropic
52.5
Fit score
Why this rank?
Coding 13Reasoning 16Cost 5Speed 7Context 12
Reasoning accounts for the largest share of this model's score.
Muse Spark 1.1
Meta
52.1
Fit score
Why this rank?
Coding 11Reasoning 12Cost 7Speed 10Context 12
Context accounts for the largest share of this model's score.
Grok 4.5
xAI
52.0
Fit score
Why this rank?
Coding 15Reasoning 16Cost 6Speed 8Context 6
Reasoning accounts for the largest share of this model's score.
Gemini 3.6 Flash
Google
51.4
Fit score
Why this rank?
Coding 11Reasoning 12Cost 8Speed 9Context 12
Context accounts for the largest share of this model's score.
GPT-5.6 Sol
OpenAI
51.1
Fit score
Why this rank?
Coding 19Reasoning 11Cost 6Speed 4Context 12
Coding accounts for the largest share of this model's score.
Muse Spark 1.2
Meta
50.9
Fit score
Why this rank?
Coding 15Reasoning 8Cost 9Speed 7Context 12
Coding accounts for the largest share of this model's score.
Qwen3.8 2.4T A95B
Alibaba
49.0
Fit score
Why this rank?
Coding 15Reasoning 8Cost 8Speed 7Context 12
Coding accounts for the largest share of this model's score.
Qwen3.8 Max (0902)
Alibaba
48.3
Fit score
Why this rank?
Coding 15Reasoning 8Cost 8Speed 6Context 12
Coding accounts for the largest share of this model's score.
Gemini 3.5 Flash
Google
47.6
Fit score
Why this rank?
Coding 9Reasoning 12Cost 5Speed 10Context 12
Context accounts for the largest share of this model's score.

GPT-6 AstraOpenAI#1 Overall MatchDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1660In: $10.00/1M tokOut: $50.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P100requested 24%applied 24%= 24.0

ReasoningRaw: No canonical data

P100requested 28%applied 28%= 28.0

ValueRaw: Out $/1M tok $50/1M tok

P0requested 20%applied 20%= 0.0

SpeedRaw: Speed (tok/s) 36 tok/s · TTFT 3,560 ms

P30requested 16%applied 16%= 4.8

ContextRaw: Context (tokens) 1.1M tokens

P100requested 12%applied 12%= 12.0

Suitability68.8

Specs →

Claude Opus 5AnthropicDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1630In: $5.00/1M tokOut: $25.00/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P100requested 24%applied 24%= 24.0

ReasoningRaw: No canonical data

P92requested 28%applied 28%= 25.7

ValueRaw: Out $/1M tok $25/1M tok

P10requested 20%applied 20%= 2.0

SpeedRaw: Speed (tok/s) 59 tok/s · TTFT 3,652 ms

P34requested 16%applied 16%= 5.4

ContextRaw: Context (tokens) 1M tokens

P96requested 12%applied 12%= 11.5

Suitability68.6

Specs →

GPT-6 SolOpenAIDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1620In: $2.00/1M tokOut: $10.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P91requested 24%applied 24%= 21.8

ReasoningRaw: No canonical data

P83requested 28%applied 28%= 23.3

ValueRaw: Out $/1M tok $10/1M tok

P24requested 20%applied 20%= 4.8

SpeedRaw: Speed (tok/s) 48 tok/s · TTFT 2,918 ms

P36requested 16%applied 16%= 5.7

ContextRaw: Context (tokens) 1.1M tokens

P100requested 12%applied 12%= 12.0

Suitability67.6

Specs →

Claude Fable 5.1AnthropicDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1620In: $10.00/1M tokOut: $50.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P91requested 24%applied 24%= 21.8

ReasoningRaw: No canonical data

P100requested 28%applied 28%= 28.0

ValueRaw: Out $/1M tok $50/1M tok

P0requested 20%applied 20%= 0.0

SpeedRaw: Speed (tok/s) 48 tok/s · TTFT 4,810 ms

P24requested 16%applied 16%= 3.8

ContextRaw: Context (tokens) 1M tokens

P96requested 12%applied 12%= 11.5

Suitability65.1

Specs →

GPT-6 LunaOpenAIDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1580In: $0.10/1M tokOut: $0.50/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P55requested 24%applied 24%= 13.1

ReasoningRaw: No canonical data

P50requested 28%applied 28%= 14.0

ValueRaw: Out $/1M tok $0.5/1M tok

P69requested 20%applied 20%= 13.8

SpeedRaw: Speed (tok/s) 62 tok/s · TTFT 2,294 ms

P46requested 16%applied 16%= 7.3

ContextRaw: Context (tokens) 1.1M tokens

P100requested 12%applied 12%= 12.0

Suitability60.2

Specs →

GPT-5.6 LunaOpenAIDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1580In: $0.20/1M tokOut: $1.20/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P55requested 24%applied 24%= 13.1

ReasoningRaw: No canonical data

P58requested 28%applied 28%= 16.3

ValueRaw: Out $/1M tok $1.2/1M tok

P56requested 20%applied 20%= 11.2

SpeedRaw: Speed (tok/s) 55 tok/s · TTFT 2,215 ms

P44requested 16%applied 16%= 7.0

ContextRaw: Context (tokens) 1.1M tokens

P100requested 12%applied 12%= 12.0

Suitability59.6

Specs →

Claude Opus 5.5AnthropicStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1660In: $4.00/1M tokOut: $20.00/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P100requested 24%applied 24%= 24.0

ReasoningRaw: No canonical data

P50requested 28%applied 28%= 13.9

ValueRaw: Out $/1M tok $20/1M tok

P16requested 20%applied 20%= 3.2

SpeedRaw: Speed (tok/s) 72 tok/s · TTFT 4,222 ms

P40requested 16%applied 16%= 6.3

ContextRaw: Context (tokens) 1M tokens

P98requested 12%applied 12%= 11.8

Suitability59.2

Specs →

DeepSeek V4.1 FlashDeepSeekStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1590In: $0.15/1M tokOut: $0.60/1M tokStreaming: 100–200 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P61requested 24%applied 24%= 14.7

ReasoningRaw: No canonical data

P29requested 28%applied 28%= 8.2

ValueRaw: Out $/1M tok $0.6/1M tok

P66requested 20%applied 20%= 13.2

SpeedRaw: Speed (tok/s) 142 tok/s · TTFT 1,000 ms

P69requested 16%applied 16%= 11.0

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability59.1

Specs →

GLM 5.3 FlashZ.aiStrong Coding

Value accounts for the largest share of this model's score.

Coding rating: 1600In: $0.15/1M tokOut: $0.50/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P66requested 24%applied 24%= 15.9

ReasoningRaw: No canonical data

P32requested 28%applied 28%= 9.0

ValueRaw: Out $/1M tok $0.5/1M tok

P82requested 20%applied 20%= 16.4

SpeedRaw: Speed (tok/s) 35 tok/s · TTFT 4,172 ms

P27requested 16%applied 16%= 4.3

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability57.6

Specs →

#10

Claude Opus 4.8AnthropicDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1600In: $5.00/1M tokOut: $25.00/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P73requested 24%applied 24%= 17.5

ReasoningRaw: No canonical data

P67requested 28%applied 28%= 18.7

ValueRaw: Out $/1M tok $25/1M tok

P10requested 20%applied 20%= 2.0

SpeedRaw: Speed (tok/s) 61 tok/s · TTFT 1,996 ms

P49requested 16%applied 16%= 7.8

ContextRaw: Context (tokens) 1M tokens

P96requested 12%applied 12%= 11.5

Suitability57.5

Specs →

#11

GPT-5.6 TerraOpenAIDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1600In: $2.00/1M tokOut: $12.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P73requested 24%applied 24%= 17.5

ReasoningRaw: No canonical data

P67requested 28%applied 28%= 18.7

ValueRaw: Out $/1M tok $12/1M tok

P21requested 20%applied 20%= 4.2

SpeedRaw: Speed (tok/s) 41 tok/s · TTFT 3,128 ms

P31requested 16%applied 16%= 5.0

ContextRaw: Context (tokens) 1.1M tokens

P100requested 12%applied 12%= 12.0

Suitability57.4

Specs →

#11

Muse Spark 1.3MetaStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1620In: $1.25/1M tokOut: $4.25/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P78requested 24%applied 24%= 18.6

ReasoningRaw: No canonical data

P38requested 28%applied 28%= 10.6

ValueRaw: Out $/1M tok $4.25/1M tok

P44requested 20%applied 20%= 8.8

SpeedRaw: Speed (tok/s) 84 tok/s · TTFT 3,500 ms

P46requested 16%applied 16%= 7.4

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability57.4

Specs →

#13

MiMo-V2.6-ProXiaomiStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1610In: $0.43/1M tokOut: $0.87/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P72requested 24%applied 24%= 17.3

ReasoningRaw: No canonical data

P38requested 28%applied 28%= 10.6

ValueRaw: Out $/1M tok $0.87/1M tok

P72requested 20%applied 20%= 14.4

SpeedRaw: Speed (tok/s) 24 tok/s · TTFT 5,839 ms

P15requested 16%applied 16%= 2.3

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability56.6

Specs →

#14

Gemini 3.7 FlashGoogleStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1590In: $0.75/1M tokOut: $3.75/1M tokStreaming: 100–200 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P61requested 24%applied 24%= 14.6

ReasoningRaw: No canonical data

P29requested 28%applied 28%= 8.2

ValueRaw: Out $/1M tok $3.75/1M tok

P46requested 20%applied 20%= 9.2

SpeedRaw: Speed (tok/s) 168 tok/s · TTFT 1,820 ms

P70requested 16%applied 16%= 11.1

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability55.1

Specs →

#15

Grok 4.6xAIStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1641In: $2.00/1M tokOut: $6.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P89requested 24%applied 24%= 21.4

ReasoningRaw: No canonical data

P35requested 28%applied 28%= 9.8

ValueRaw: Out $/1M tok $6/1M tok

P38requested 20%applied 20%= 7.6

SpeedRaw: Speed (tok/s) 48 tok/s · TTFT 1,750 ms

P47requested 16%applied 16%= 7.4

ContextRaw: Context (tokens) 500k tokens

P64requested 12%applied 12%= 7.7

Suitability53.9

Specs →

#16

Grok 4.7xAIStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1610In: $1.60/1M tokOut: $4.80/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P72requested 24%applied 24%= 17.3

ReasoningRaw: No canonical data

P38requested 28%applied 28%= 10.6

ValueRaw: Out $/1M tok $4.8/1M tok

P42requested 20%applied 20%= 8.4

SpeedRaw: Speed (tok/s) 61 tok/s · TTFT 1,288 ms

P56requested 16%applied 16%= 9.0

ContextRaw: Context (tokens) 500k tokens

P64requested 12%applied 12%= 7.7

Suitability53.0

Specs →

#17

Claude Sonnet 5AnthropicDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1580In: $2.00/1M tokOut: $10.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P55requested 24%applied 24%= 13.1

ReasoningRaw: No canonical data

P58requested 28%applied 28%= 16.3

ValueRaw: Out $/1M tok $10/1M tok

P24requested 20%applied 20%= 4.8

SpeedRaw: Speed (tok/s) 49 tok/s · TTFT 2,193 ms

P43requested 16%applied 16%= 6.8

ContextRaw: Context (tokens) 1M tokens

P96requested 12%applied 12%= 11.5

Suitability52.5

Specs →

#18

Muse Spark 1.1MetaDeep Reasoning

Context accounts for the largest share of this model's score.

Coding rating: 1570In: $1.25/1M tokOut: $4.25/1M tokStreaming: 200+ tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P45requested 24%applied 24%= 10.9

ReasoningRaw: No canonical data

P42requested 28%applied 28%= 11.7

ValueRaw: Out $/1M tok $4.25/1M tok

P37requested 20%applied 20%= 7.4

SpeedRaw: Speed (tok/s) 233 tok/s · TTFT 2,806 ms

P63requested 16%applied 16%= 10.1

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability52.1

Specs →

#19

Grok 4.5xAIDeep Reasoning

Reasoning accounts for the largest share of this model's score.

Coding rating: 1590In: $2.00/1M tokOut: $6.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P64requested 24%applied 24%= 15.3

ReasoningRaw: No canonical data

P58requested 28%applied 28%= 16.3

ValueRaw: Out $/1M tok $6/1M tok

P32requested 20%applied 20%= 6.4

SpeedRaw: Speed (tok/s) 47 tok/s · TTFT 1,412 ms

P53requested 16%applied 16%= 8.4

ContextRaw: Context (tokens) 500k tokens

P47requested 12%applied 12%= 5.6

Suitability52.0

Specs →

#20

Gemini 3.6 FlashGoogleDeep Reasoning

Context accounts for the largest share of this model's score.

Coding rating: 1570In: $0.75/1M tokOut: $3.75/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P45requested 24%applied 24%= 10.9

ReasoningRaw: No canonical data

P42requested 28%applied 28%= 11.7

ValueRaw: Out $/1M tok $3.75/1M tok

P39requested 20%applied 20%= 7.8

SpeedRaw: Speed (tok/s) 53 tok/s · TTFT 1,290 ms

P57requested 16%applied 16%= 9.0

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability51.4

Specs →

#21

GPT-5.6 SolOpenAIStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1620In: $2.00/1M tokOut: $10.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P78requested 24%applied 24%= 18.6

ReasoningRaw: No canonical data

P38requested 28%applied 28%= 10.6

ValueRaw: Out $/1M tok $10/1M tok

P29requested 20%applied 20%= 5.8

SpeedRaw: Speed (tok/s) 35 tok/s · TTFT 4,621 ms

P26requested 16%applied 16%= 4.1

ContextRaw: Context (tokens) 1.1M tokens

P100requested 12%applied 12%= 12.0

Suitability51.1

Specs →

#22

Muse Spark 1.2MetaStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1590In: $1.25/1M tokOut: $4.25/1M tokStreaming: 100–200 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P61requested 24%applied 24%= 14.6

ReasoningRaw: No canonical data

P29requested 28%applied 28%= 8.2

ValueRaw: Out $/1M tok $4.25/1M tok

P44requested 20%applied 20%= 8.8

SpeedRaw: Speed (tok/s) 100 tok/s · TTFT 4,308 ms

P46requested 16%applied 16%= 7.3

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability50.9

Specs →

#23

Qwen3.8 2.4T A95BAlibabaStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1590In: $2.00/1M tokOut: $6.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P61requested 24%applied 24%= 14.6

ReasoningRaw: No canonical data

P29requested 28%applied 28%= 8.2

ValueRaw: Out $/1M tok $6/1M tok

P38requested 20%applied 20%= 7.6

SpeedRaw: Speed (tok/s) 30 tok/s · TTFT 1,410 ms

P42requested 16%applied 16%= 6.6

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability49.0

Specs →

#24

Qwen3.8 Max (0902)AlibabaStrong Coding

Coding accounts for the largest share of this model's score.

Coding rating: 1590In: $2.00/1M tokOut: $6.00/1M tokStreaming: <50 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P61requested 24%applied 24%= 14.6

ReasoningRaw: No canonical data

P29requested 28%applied 28%= 8.2

ValueRaw: Out $/1M tok $6/1M tok

P38requested 20%applied 20%= 7.6

SpeedRaw: Speed (tok/s) 37 tok/s · TTFT 2,240 ms

P38requested 16%applied 16%= 6.1

ContextRaw: Context (tokens) 1M tokens

P98requested 12%applied 12%= 11.8

Suitability48.3

Specs →

#25

Gemini 3.5 FlashGoogleDeep Reasoning

Context accounts for the largest share of this model's score.

Coding rating: 1560In: $1.50/1M tokOut: $9.00/1M tokStreaming: 50–100 tok/s

Why this rank?

Raw benchmark → normalized percentile → weight → contribution · every requested dimension has published evidence

CodingRaw: No canonical data

P36requested 24%applied 24%= 8.7

ReasoningRaw: No canonical data

P42requested 28%applied 28%= 11.7

ValueRaw: Out $/1M tok $9/1M tok

P26requested 20%applied 20%= 5.2

SpeedRaw: Speed (tok/s) 78 tok/s · TTFT 1,332 ms

P63requested 16%applied 16%= 10.0

ContextRaw: Context (tokens) 1M tokens

P100requested 12%applied 12%= 12.0

Suitability47.6

Specs →

Per Page:

U = published provisionally after admin review; final CompareLLM verification is pending.

Platform Discovery & Next Steps

Explore More AI Intelligence Tools

Continue exploring independent model comparisons, hardware fit calculators, and live market movements.

Back to Homepage Overview

Head-to-Head

Pairwise Model Comparisons

Compare any two models on our category ratings, speed classes and token pricing.

Compare Models

Ranked Picks

Curated 'Best-Of' Indexes

Targeted rankings for Best Coding LLMs, Best Cheap APIs, Shortest Wait, and top-tier models.

Browse Categories

50+ DGX, Mac & GPUs

AI Hardware Calculator

Interactive VRAM calculator, quantization levels (FP16, Q8, Q4), KV cache context, and local hardware fit.

Calculate VRAM

Daily Audit

Live Market Intelligence

Live audit trail of benchmark updates, new model releases, and API price cuts.

View Changes

Smart Wizard

Interactive Model Finder

Answer 5 quick questions to compute deterministic model recommendations for your use case.

Launch Finder

Data Integrity

Methodology & Standards

How a CompareLLM rating is computed, what evidence it admits, and how prices and speed classes are recorded.

Read Methodology