Why is this comparison indexable?

It has 6 verifiable comparison points, and both models have pricing or benchmark data.

Are missing metrics invented?

No. Metrics without LMSpeed data are omitted from this report.

Back to models

Data points: 100

Model compare

Claude Sonnet 4.6 vs Phi 4

The readout for Claude Sonnet 4.6 and Phi 4, before the detailed comparison sheet.

Model A

Claude Sonnet 4.6

Anthropic

Leading

vs

Model B

Phi 4

Microsoft

Contender

Key Takeaways

Weighted outcome: Claude Sonnet 4.6. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.

Decision read

Claude Sonnet 4.6

Claude Sonnet 4.6 has the higher weighted result; Model A / B score 96 to 4.

Evidence depth

100 data points

Includes 17 benchmark rows, 0 audit samples, and 6 provider examples.

Selection signal

Start with Claude Sonnet 4.6

The charts below split 23 high-signal samples across speed, scores, and audit health.

Change comparison

Switch either side of this report to compare another model with the same LMSpeed data pipeline.

Model AModel B

Comparison sheet

This report only uses LMSpeed data for Claude Sonnet 4.6 and Phi 4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.

Model compare	Claude Sonnet 4.6	Phi 4
Overall leader	Leading	Contender
Weighted overall score	96.0 pts	4.0 pts
Benchmark category leads	6 categories	0 categories
Operational advantages	Cheapest input price, Average speed, Free providers, Provider coverage	First-token latency
Context window	1M tokens	16.4K tokens
Max output	128K tokens	16.4K tokens
Modalities	Input TextImageFile Output Text

The overall result weights benchmark capability categories at 80% and price, API speed/latency, and availability at 20%. Recent test volume does not affect the winner, and missing benchmark categories are excluded.

Model metadata

Model compare	Claude Sonnet 4.6	Phi 4
Developer	Anthropic	Microsoft
Released	Feb 2026	Jan 2025
Parameters	No data	No data
Tokenizer	Claude	Other
Knowledge cutoff	No data	2024-06-30
OpenRouter ID	anthropic/claude-sonnet-4.6	microsoft/phi-4
References	No data	No data

When to choose each model

This report only uses LMSpeed data for Claude Sonnet 4.6 and Phi 4: pricing, speed aggregates, third-party benchmark scores, and shared provider samples.

Claude Sonnet 4.6

Claude Sonnet 4.6 is stronger in benchmark categories (Agents, Coding, Reasoning, Knowledge, Math) and operational dimensions (Cheapest input price, Average speed, Free providers, Provider coverage).

Phi 4

Phi 4 has these operational advantages: First-token latency.

Benchmark score comparison

Third-party benchmark profile synced into LMSpeed; only metrics available for both models are shown.

Category performance

Compare benchmark category scores on a 0-100 scale. Select a category to inspect the gap.

Model A coverage: 7 / 8
Model B coverage: 6 / 8
Shared: 6 shared categories

Avg. score

Claude Sonnet 4.6

53.2

Avg. score

Phi 4

38.4

Agents

Claude Sonnet 4.6 leads by 24.6

Claude Sonnet 4.653.6

Phi 429

Coding

Claude Sonnet 4.6 leads by 16.7

Claude Sonnet 4.656.5

Phi 439.8

Reasoning

Claude Sonnet 4.6 leads by 10.8

Claude Sonnet 4.650.3

Phi 439.5

Knowledge

Claude Sonnet 4.6 leads by 32.6

Claude Sonnet 4.670.6

Phi 438

Math

Claude Sonnet 4.6 leads by 6.4

Claude Sonnet 4.653.2

Phi 446.8

Multilingual

No data

Claude Sonnet 4.6-

Phi 4-

Multimodal

Claude Sonnet 4.6

Claude Sonnet 4.644.5

Phi 4-

Instruction following

Claude Sonnet 4.6 leads by 6.3

Claude Sonnet 4.643.8

Phi 437.5

Professional benchmark details

Metric-level scores with benchmark source, rank depth, confidence, error, and evaluation date where available.

Group

Performanceverified

Time to first answer token

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

73.67 s

Rank #129/139 · confidence 4

+71.54 s

Phi 4

winner

2.14 s

Rank #57/139 · confidence 4

Performanceverified

Output speed

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

winner

74.6 tok/s

Rank #83/139 · confidence 4

+45.7 tok/s

Phi 4

28.9 tok/s

Rank #135/139 · confidence 4

Performanceverified

Time to first token

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

73.67 s

Rank #133/139 · confidence 4

+71.54 s

Phi 4

winner

2.14 s

Rank #102/139 · confidence 4

Pricingverified

Input price

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

$3.00/M

Rank #139/163 · confidence 4

+$2.88/M

Phi 4

winner

$0.125/M

Rank #17/163 · confidence 4

Pricingverified

Blended price

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

$6.00/M

Rank #139/163 · confidence 4

+$5.78/M

Phi 4

winner

$0.219/M

Rank #19/163 · confidence 4

Pricingverified

Output price

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

$15.00/M

Rank #138/163 · confidence 4

+$14.50/M

Phi 4

winner

$0.500/M

Rank #21/163 · confidence 4

Reasoningverified

GPQA

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

winner

87.5%

Rank #30/187 · confidence 4

+30.0%

Phi 4

57.5%

Rank #150/187 · confidence 4

Reasoningverified

HLE

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

winner

30.0%

Rank #30/186 · confidence 4

+25.9%

Phi 4

4.1%

Rank #156/186 · confidence 4

Reasoningreported

MMLU-Pro

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

79.2%

Rank #68/125 · confidence 3 · eval date 2026-02-01

+7.8%

Phi 4

71.4%

Rank #98/125 · confidence 4

Codingverified

SciCode

SourceClaude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Claude Sonnet 4.6

winner

46.8%

Rank #28/184 · confidence 4

+20.8%

Phi 4

26.0%

Rank #154/184 · confidence 4

Codingreported

AA-SciCode

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

46.9

Rank #33/96 · confidence 3 · eval date 2026-02-01

+20.9

Phi 4

26.0

Rank #90/96 · confidence 1 · eval date 2025-01-01

Knowledgereported

AA-Omniscience Accuracy

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

38.0

Rank #29/90 · confidence 3 · eval date 2026-02-01

+24.8

Phi 4

13.2

Rank #89/90 · confidence 1 · eval date 2025-01-01

Knowledgereported

AA-Omniscience Hallucination Rate

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

65.9

Rank #58/90 · confidence 3 · eval date 2026-02-01

+14.6

Phi 4

winner

80.5

Rank #36/90 · confidence 1 · eval date 2025-01-01

Knowledgereported

Artificial Analysis Intelligence Index

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

35.9

Rank #43/99 · confidence 3 · eval date 2026-02-01

+31.0

Phi 4

4.9

Rank #98/99 · confidence 1 · eval date 2025-01-01

Knowledgereported

AA-GPQA Diamond

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

79.9

Rank #63/96 · confidence 3 · eval date 2026-02-01

+22.4

Phi 4

57.5

Rank #89/96 · confidence 1 · eval date 2025-01-01

Knowledgereported

AA-HLE

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

13.2

Rank #63/96 · confidence 3 · eval date 2026-02-01

+9.1

Phi 4

4.1

Rank #87/96 · confidence 1 · eval date 2025-01-01

Instruction followingreported

AA-IFBench

SourceClaude Sonnet 4.6

Claude Sonnet 4.6

winner

41.2

Rank #69/86 · confidence 3 · eval date 2026-02-01

+17.7

Phi 4

23.5

Rank #85/86 · confidence 1 · eval date 2025-01-01

API audit comparison

Latest completed audits from shared providers, with four safety and integrity score groups plus report links.

Provider	Claude Sonnet 4.6	Phi 4
No completed audits are available from shared providers yet.

Provider examples

Speed aggregates and input/output pricing share each provider row for real API selection and migration cost checks.

Provider	Claude Sonnet 4.6	Phi 4
10 tests	Claude Sonnet 4.6 speed / latency N/A / N/A input / output No data	Phi 4 speed / latency 43 tok/s / 1320ms input / output No data
10 tests	Claude Sonnet 4.6 claude-sonnet-4.6 speed / latency 46 tok/s / 6386ms input / output $0.100/request	Phi 4 phi-4 speed / latency N/A / N/A input / output $0.010/request
0 tests	Claude Sonnet 4.6 speed / latency N/A / N/A input / output No data	Phi 4 speed / latency N/A / N/A input / output No data
0 tests	Claude Sonnet 4.6 speed / latency N/A / N/A input / output No data	Phi 4 speed / latency N/A / N/A input / output No data
0 tests	Claude Sonnet 4.6 speed / latency N/A / N/A input / output No data	Phi 4 speed / latency N/A / N/A input / output No data
	Claude Sonnet 4.6 claude-sonnet-4-6 speed / latency No data input / output $0.021/request	Phi 4 phi-4 speed / latency No data input / output $57.53/M

FAQ

Weighted outcome: Claude Sonnet 4.6. Benchmark capability categories carry 80%, while price, API performance, and availability carry 20%.

Why is this comparison indexable?: It has 6 verifiable comparison points, and both models have pricing or benchmark data.
Are missing metrics invented?: No. Metrics without LMSpeed data are omitted from this report.

Input

Text

Output

Text

Features	Text inputImage inputFile inputText outputTool callingStructured outputsJSON modeReasoning	Text inputText outputStructured outputsJSON mode

/

$57.53/M

Comparison sheet

Model metadata

When to choose each model

Benchmark score comparison

Category performance

Agents

Coding

Reasoning

Knowledge

Math

Multilingual

Multimodal

Instruction following

Professional benchmark details

Time to first answer token

Output speed

Time to first token

Input price

Blended price

Output price

GPQA

HLE

MMLU-Pro

SciCode

AA-SciCode

AA-Omniscience Accuracy

AA-Omniscience Hallucination Rate

Artificial Analysis Intelligence Index

AA-GPQA Diamond

AA-HLE

AA-IFBench

API audit comparison

Provider examples

FAQ

Related compare reports